Content uploaded by Jorge G. Álvarez-Romero
Author content
All content in this area was uploaded by Jorge G. Álvarez-Romero on Dec 04, 2015
Content may be subject to copyright.
Assessment of the relative risk of
degraded water quality to ecosystems of
the Great Barrier Reef
Lead Authors: Jon Brodie, Jane Waterhouse, Jeffrey Maynard
Contributors: John Bennett, Miles Furnas, Michelle Devlin, Stephen Lewis,
Catherine Collier, Britta Schaffelke, Katharina Fabricius, Caroline Petus, Eduardo da
Silva, Daniel Zeh, Lucy Randall, Vittorio Brando, Len McKenzie, Dominique O’Brien,
Rachael Smith, Michael Warne, Richard Brinkman, Hemerson Tonin, Zoe
Bainbridge, Rebecca Bartley, Andrew Negri, Ryan Turner, Aaron Davis, Christie
Bentley, Jochen Mueller, Jorge Alvarez-Romero, Nyssa Henry, David Waters, Hugh
Yorkston, Dieter Tracey
Report No. 13/28
November 2013
Assessment of the relative risk of degraded water
quality to ecosystems of the Great Barrier Reef
A Report for the Queensland Department of the Environment and Heritage
Protection
Report No. 13/28
November 2013
Lead by Jon Brodie, Jane Waterhouse and Jeffrey Maynard with many
contributing authors
Centre for Tropical Water & Aquatic Ecosystem Research
(TropWATER)
James Cook University
Townsville
Phone : (07) 4781 4262
Email: TropWATER@jcu.edu.au
Web: www.jcu.edu.au/tropwater/
i
Lead Authors:
Jon Brodie
TropWATER James Cook University
Jane Waterhouse
TropWATER James Cook University and C2O Consulting
Jeffrey Maynard
Laboratoire d’Excellence 'CORAIL' USR 3278 CNRS – EPHE, CRIOBE, Papetoai,
Moorea, Polynesie Francaise and C2O Consulting
Contributors:
John Bennett
Department of Environment and Heritage Protection
Miles Furnas
Australian Institute of Marine Science
Michelle Devlin
TropWATER James Cook University and C2O Consulting
Stephen Lewis
TropWATER James Cook University
Catherine Collier
James Cook University
Britta Schaffelke
Australian Institute of Marine Science
Katharina Fabricius
Australian Institute of Marine Science
Caroline Petus
TropWATER James Cook University
Eduardo Da Silva
TropWATER James Cook University
Daniel Zeh
James Cook University
Lucy Randall
Australian Bureau of Agricultural and Resource Economics and Sciences
Vittorio Brando
Commonwealth Science and Industrial Research Organisation
Len McKenzie
TropWATER James Cook University
Dominique O’Brien
TropWATER James Cook University
Rachael Smith
Department of Science, Information Technology, Innovation and the Arts
Michael Warne
Department of Science, Information Technology, Innovation and the Arts
Richard Brinkman
Australian Institute of Marine Science
Hemerson Tonin
Australian Institute of Marine Science
Zoe Bainbridge
TropWATER James Cook University
Rebecca Bartley
Australian Institute of Marine Science
Andrew Negri
Australian Institute of Marine Science
Ryan Turner
Department of Science, Information Technology, Innovation and the Arts
Aaron Davis
TropWATER James Cook University
Christie Bentley
EnTOX, University of Queensland
Jochen Mueller
EnTOX, University of Queensland
Jorge G. Alvarez-Romero
TropWATER James Cook University and Australian Research Council Centre of
Excellence for Coral Reef Studies, James Cook University
Nyssa Henry
Department of Science, Information Technology, Innovation and the Arts
David Waters
Department of Natural Mines and Resources
Hugh Yorkston
Great Barrier Reef Marine Park Authority
Dieter Tracey
Green Ant Photo Design
ii
Citation
This report should be cited as:
Brodie, J., Waterhouse, J., Maynard, J., Bennett, J., Furnas, M., Devlin, M., Lewis, S., Collier, C., Schaffelke, B.,
Fabricius, K., Petus, C., da Silva, E., Zeh, D., Randall, L., Brando, V., McKenzie, L., O’Brien, D., Smith, R., Warne,
M.St.J., Brinkman, R., Tonin, H., Bainbridge, Z., Bartley, R., Negri, A., Turner, R.D.R., Davis, A., Bentley, C.,
Mueller, J., Alvarez-Romero, J.G., Henry, N., Waters, D., Yorkston, H., Tracey, D., 2013. Assessment of the relative
risk of water quality to ecosystems of the Great Barrier Reef. A report to the Department of the Environment and
Heritage Protection, Queensland Government, Brisbane. TropWATER Report 13/28, Townsville, Australia.
For further information contact:
Jon Brodie
Centre for Tropical Water & Aquatic Ecosystem Research (TropWATER)
James Cook University
Email: Jon.Brodie@jcu.edu.au
This publication has been compiled by the Centre for Tropical Water & Aquatic Ecosystem Research
(TropWATER), James Cook University.
© James Cook University, 2013.
Except as permitted by the Copyright Act 1968, no part of the work may in any form or by any electronic,
mechanical, photocopying, recording, or any other means be reproduced, stored in a retrieval system or be
broadcast or transmitted without the prior written permission of TropWATER. The information contained herein
is subject to change without notice. The copyright owner shall not be liable for technical or other errors or
omissions contained herein. The reader/user accepts all risks and responsibility for losses, damages, costs and
other consequences resulting directly or indirectly from using this information.
Enquiries about reproduction, including downloading or printing the web version, should be directed to
jane.waterhouse@jcu.edu.au.
iii
Context to the report
The Queensland Government’s Reef Water Quality Program provided funding for this up-to-date assessment of
the relative risks of degraded water quality to ecosystems of the Great Barrier Reef to inform the development
of Reef Plan 2013. The assessment, and its supplementary studies, informed the development of the 2013
Scientific Consensus Statement and, in particular, formed the basis of Chapter 3 of that statement. The
Scientific Consensus Statement underpinned the management strategies and their priorities in Reef Plan 2013,
including setting investment priorities for the Australian Government’s Reef Rescue 2 program. This report, its
supporting studies and the information collated for this assessment, will be a “living resource” that will continue
to inform discussion of Reef Plan and other related management, investment and monitoring priorities at
regional level in the reef catchments as their Water Quality Improvement Plans are updated.
The Queensland Government is committed to ongoing investment in Reef Plan science and, as new information
becomes available, government policy will be adapted to take account of it. The report is provided in good faith,
on the understanding that the information is not used out of the context explained above.
Acknowledgements
The authors of this report would like to thank the Queensland Department of Environment and Heritage
Protection for the project funding.
A number of individuals have also assisted with data acquisition and input including Alana Grech and Rob Coles
for advice on seagrass data, Dieter Tracey for assistance with graphics, John Bennett for project management
and guidance, and Lucy Randall for training with MCAS-S.
Special thanks also go to Professor Barry Hart and Dr Jeff Dambacher for their comprehensive review of this
report; their input and suggestions have contributed to a much improved final version of the report. The authors
would also like to acknowledge Dr Roger Shaw and Mike Grundy for their valuable input and review of the
associated Reef Plan Science Consensus Statement chapter on the relative risks of degraded water quality to the
Great Barrier Reef which also informed the revision of this report.
Disclaimers
TropWATER advises that the information contained in this publication comprises general statements based on
scientific research. The reader is advised and needs to be aware that such information may be incomplete or
unable to be used in any specific situation. To the extent permitted by law, TropWATER (including its employees
and consultants) excludes all liability to any person for any consequences, including but not limited to all losses,
damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication
(in part or in whole) and any information or material contained in it.
iv
Table of Contents
Citation ........................................................................................................................................................................ i
Acknowledgements ................................................................................................................................................... iii
Disclaimers ................................................................................................................................................................ iii
Executive Summary ....................................................................................................................................................1
1 Introduction ..................................................................................................................................................... 11
Part A: Assessment of the relative risk of degraded water quality to ecosystems of the GBR ............................... 18
2 Assessment of the risk of pollutants to ecosystems of the Great Barrier Reef including differential risk
between sediments, nutrients and pesticides, and among NRM regions .............................................................. 18
2.1 Summary of findings ................................................................................................................................. 18
2.2 Introduction .............................................................................................................................................. 20
2.2.1 Risk assessment framework ................................................................................................................ 20
2.3 Methods .................................................................................................................................................... 25
2.3.1 Selecting and classifying variables ....................................................................................................... 25
a) Exceedance of suspended solids concentration thresholds ............................................................ 32
b) Chlorophyll concentration exceedance ........................................................................................... 35
c) Pollutant loading in river plumes .................................................................................................... 37
d) Pesticide concentrations ................................................................................................................. 40
e) Recognising and assessing uncertainties in the selection of variables ........................................... 42
2.3.2 Estimating habitat area ....................................................................................................................... 43
2.3.3 Assessment Method - Part 1: Differential risk between pollutants on GBR ecosystems ................... 43
2.3.4 Assessment Method - Part 2: Combined risk of degraded water quality to GBR ecosystems ............ 44
2.3.5 Assessment Method - Part 3: Relative risk of degraded water quality to GBR ecosystems ............... 48
2.4 Results ....................................................................................................................................................... 49
2.4.1 Habitat and regional NRM areas ......................................................................................................... 49
2.4.2 Part 1: Relative importance of different pollutants to GBR ecosystems............................................. 49
a) Sediments ........................................................................................................................................ 51
b) Nutrients .......................................................................................................................................... 60
c) Pesticides ......................................................................................................................................... 68
d) Comparisons among variables and Regions .................................................................................... 73
2.4.3 Part 2: Combined risk of degraded water quality to GBR ecosystems ................................................ 76
2.4.4 Part 3: Relative risk of degraded water quality to GBR ecosystems ................................................... 81
a) Assessment of end-of-catchment pollutant loads .......................................................................... 81
b) Combined assessment: Relative Risk Index ..................................................................................... 84
2.4.5 Conclusions .......................................................................................................................................... 87
v
Part B: Overall conclusions of the relative risk between sediments, nutrients and pesticides and between land
uses, industries and catchments in the GBR ........................................................................................................... 90
3 Relative risk from degraded water quality to GBR ecosystems ....................................................................... 90
3.1 Additional information to support the assessment in Part A ................................................................... 90
3.2 Overall assessment of the relative risk of degraded water quality to GBR ecosystems .......................... 94
3.3 Overall conclusions ................................................................................................................................... 98
4 Limitations to the risk assessment and future improvements ...................................................................... 100
5 References ...................................................................................................................................................... 107
Appendix 1. Further information regarding remote sensing assessments ........................................................... 121
Appendix 2. Sensitivity analysis of options for combining water quality variables for determining relative risk 124
Page 1
Executive Summary
A risk assessment method was developed and applied to the Great Barrier Reef (GBR) to provide robust and
scientifically defensible information for policy makers and catchment managers on the key land-based pollutants
of greatest risk to the health of the two main GBR ecosystems (coral reefs and seagrass beds). This information
was used to inform management prioritisation for Reef Rescue 2 and Reef Plan 3. The risk assessment method
needed to take account of the fact that catchment-associated risk will vary with distance from the river mouth,
with coastal habitats nearest to river mouths most impacted by poor marine water quality.
The main water quality pollutants of concern for the GBR are enhanced levels of suspended sediments, excess
nutrients and pesticides added to the GBR lagoon from the adjacent catchments. Until recently, there has been
insufficient knowledge about the relative exposure to and effects of these pollutants to guide effective
prioritisation of the management of their sources.
This report is presented in 2 main parts: Part A describes the risk assessment method and reports the main
findings; Part B provides the overall conclusions which are also informed by a number of supporting studies
which are provided as a separate report (Waterhouse, 2013). The key findings of all parts of the report and
supporting studies are summarised below.
Part A: Risk assessment
The risk assessment method is described in Part A of this report and used a combination of qualitative and semi-
quantitative information about the influence of individual catchments in the 6 natural resource management
(NRM) regions on coral reefs and seagrass ecosystems. The marine boundaries used for each NRM region are
those accepted officially by the Great Barrier Reef Marine Park Authority. The area of GBR lagoon waters within
each marine NRM region were also reported in recognition of other important habitats and populations that
exist in these areas.
The relative risk of degraded water quality among NRM regions was determined by combining information on
end-of-catchment pollutant loads and the estimated ecological risk of water quality to coral reefs and seagrass
meadows in the GBR. Modelled end-of-catchment pollutant loads (generated from the Source Catchments
model framework for the Paddock to Reef Program) were obtained for each NRM region for key pollutants, and
only the anthropogenic portions of total pollutant loads were considered in relating the relative risk to NRM
regions. The anthropogenic load is calculated as the difference between the long term average annual load, and
the estimated pre-European annual load. This information was used to define a ‘Loads Index’.
Ecological risk is generally defined as the product of the likelihood of an effect occurring and the consequences if
that effect was to occur. However, in this assessment there is some inconsistency in our capacity across the
variables to produce a true likelihood or true consequence estimate as mostly we have no or limited ability to
produce these estimates right now. Therefore, ecological risk in the GBR is expressed simply as the area of coral
reefs and seagrass meadows within a range of assessment classes (very low to very high relative risk) for several
water quality variables in each NRM region in the GBR lagoon. Our method for calculating risk essentially
assesses the likelihood of exceedance of a selected threshold. This likelihood was set as 1 for a parameter and
location if observations or modelled data indicate that the threshold was exceeded. Conversely, the likelihood
was set as 0 if observations or modelled data indicate that the threshold was not exceeded. As consequences
are mostly unknown at a regional or species level, potential impact was calculated as the area of coral reef,
seagrass meadows and area of GBR lagoon waters (in km2) within the highest assessment classes of the water
quality variables (reflecting the highest severity of influence). The effects of multiplying the habitat area by 1 or
0 for the likelihood mean that the final assessment of risk in this assessment is only an indication of potential
Page 2
impact - the area of coral reef and seagrass meadows in which exceedance of an agreed threshold was modeled
or observed. This becomes an assessment of ‘relative risk’ by comparing the areas of each habitat affected by
the highest assessment classes of the variables among NRM regions, and was used to generate a ‘Marine Risk
Index’ for coral reefs and seagrass meadows.
A suite of water quality variables were chosen that represent the pollutants of greatest concern with regards to
agricultural runoff and potential impacts on GBR ecosystems. The selection of variables was informed by the
supporting studies described in Part B of this report (Waterhouse, 2013), and include exceedance of ecologically-
relevant thresholds for concentrations of total suspended solids (TSS) and chlorophyll a obtained from daily
remote sensing observations, and the distribution of key pollutants including TSS, dissolved inorganic nitrogen
(DIN) and photosystem II-inhibiting herbicides (PSII herbicides) in the marine environment during flood
conditions (based on end-of-catchment loads and plume loading estimates). A spatial variable was included that
represents the area of the GBR lagoon where primary crown-of-thorns starfish (COTS) outbreaks have been
observed. COTS outbreaks are an important cause of coral loss on the GBR and appear to be a response to
excess nutrient runoff from certain catchments that impact this ‘COTS initiation zone’. In recognition of the
importance of the influence of catchment discharges in driving COTS outbreaks, an index of regional
contributions of river discharges to the COTS initiation zone is also included for coral reefs (COTS Influence
Index).
The three indexes were then combined to generate a Relative Risk Index, which ultimately ranks the relative risk
of degraded water quality to coral reefs and seagrass meadows in the GBR among NRM regions.
Supporting Studies (see chapters within Waterhouse, 2013)
The supporting studies of the project informed the selection of variables and methods of analysis used in the risk
assessment. The studies also strengthen our understanding of the consequences of pollutant impacts on coral
reefs and seagrasses. In particular: Chapter 1 emphasises the importance of nutrients in the initiation of COTS
primary outbreaks and therefore loss of coral cover; Chapter 2 confirms the relative importance between
nitrogen and phosphorus in driving productivity in the GBR lagoon; Chapter 3 reviews the effects of sediments
and sedimentation on coral reef communities; Chapter 4 provides new information on the relative risk of PSII
herbicides to coastal and marine ecosystems, identifying the highest risk areas to be in freshwater and coastal
wetlands, estuarine areas and coastal seagrass and inshore reef communities; Chapter 5 develops relationships
between flood plume frequency and end of catchment pollutant loadings to define plume water types related to
water quality characteristics; Chapter 6 provides a case study that shows a correlation between plume water
types and seagrass cover with a supporting review on the impacts of water quality on seagrass in an appendix
(Chapter 7); and Chapter 8 summarises the results of water quality and phytoplankton samples collected in
variable flood plume conditions between 2010 and 2012 to improve our understanding of phytoplankton
population dynamics.
Key findings of the Supporting Studies include:
A variety of experimental, modelling and observational evidence circumstantially, but strongly, supports the
hypothesis that COTS primary outbreaks are initiated by an episode of greatly enhanced larval survival
during conditions producing increased food availability for the filter-feeding pelagic larval stages (Chapter 1).
The four primary COTS outbreaks originating in the Lizard Island - Cairns region (14.5-17°S) since 1960 follow
2-5 years after wet seasons when early season (November-February) aggregate discharges from the
Burdekin to Daintree Rivers exceeded 10 Km3.
Hydrodynamic modelling and estimates of DIN loads were used to rank the individual contribution of
significant rivers between the Daintree (16°S) and Burdekin (19°S) to regional runoff influences on the Lizard
Page 3
Island – Cairns initiation region (Chapter 1). The initiation region is divided into two areas – north and south
of Undine Reef (16°S). On a runoff volumetric basis, the Daintree has the largest influence followed by the
Russell-Mulgrave and Tully. With DIN loads included, the Russell-Mulgrave and Tully Rivers are the most
important in the northern part of the initiation region (14.5-16°S). For the southern sub-region (16-17°S) of
the initiation region the Daintree, Barron and Burdekin Rivers ranked highest, indicating the additional
significant influence of northward transport of the Burdekin River plume. When DIN loads are included, the
influence of the Burdekin increases greatly, particularly in the southern part of the initiation region (16-
17°S).
Dissolved inorganic and particulate forms of nutrients discharged into the GBR are both important in driving
ecological effects but increased nitrogen inputs are more important than phosphorus inputs (Chapter 2).
Dissolved inorganic forms of nitrogen and phosphorus are considered to be of greatest concern compared to
dissolved organic and particulate forms of nutrients, as they are immediately and completely bioavailable
for algal growth. Particulate forms mostly become bioavailable over longer time frames, and dissolved
organic forms typically have limited and delayed bioavailability.
Across the 35 basins in the GBR catchment area, the highest annual average loads of total suspended solids
are derived from the Burdekin River (3,306 kt/year), Fitzroy River (1,805 kt/year), Herbert River (482
kt/year), Mary River (362 kt/year) and Don River (330 kt/year) (Chapter 3). Recent research has highlighted
gully and streambank erosion as the dominant sediment erosion processes occurring in the larger dry
tropical river catchments of the GBR and Gulf of Carpentaria. Prior to these field studies, catchment
modelling had identified hillslope erosion as the dominant source of sediment in these catchments. The very
fine sediment fraction (<10 µm) has a shorter residence time than coarser sediment, and is less likely to be
captured in storage pathways, such as reservoirs.
Floodplume studies have found most sediment exported from GBR rivers (e.g. Burdekin, Fitzroy, Tully and
Burnett Rivers and the Mackay Whitsunday rivers) is deposited within close proximity of the river
mouth/estuary, near-shore zone and inner shelf of the GBR, with potential for remobilisation during
subsequent wind and tide driven resuspension events. Such resuspension events can result in higher
turbidity levels than measured in initial flood plumes. The delivery of this new sediment to the inner shelf
sediment wedge plays a critical role on inshore turbidity regimes, with a recent study highlighting the
influence of increased river flow/sediment loads on temporal variation in inshore turbidity.
Turbidity reduces light for benthic organisms such as seagrass and corals (Chapter 3). Coastal coral reefs do
grow in turbid water conditions at shallow depths, however biodiversity declines as a function of increased
turbidity throughout the GBR. Variability in light conditions may be a greater stress than chronically reduced
light, as energy demands to adjust to varying light conditions result in suboptimal energy gains.
Sediment properties strongly influence the effects of sedimentation on corals. Nutrient enriched fine
terrestrial silts found along the inshore GBR are most detrimental. Early life stages (e.g. fertilisation,
settlement) of corals are particularly susceptible to poor water quality and sedimentation.
Coral reefs most vulnerable to damage from turbidity and sedimentation are those found in locations with
weak currents such as embayments (e.g. Keppel Bay, Cleveland Bay, Missionary Bay) and sheltered zones on
deeper reef slopes, in places where fish abundances are low, and in regions that are frequently affected by
other forms of disturbance such as cyclones, bleaching or crown-of-thorns starfish predation.
Seagrass meadows are widespread and ecologically critical parts of the Great Barrier Reef Marine Park
(Chapter 7). Recent, widespread loss of seagrass has conclusively demonstrated their sensitivity to water
Page 4
quality, and the ecological ramifications of seagrass loss including dugong and turtle mortality. Seagrass
meadow distribution, abundance, productivity, composition and resilience are structured by chronic and
acute water quality impacts.
Both suspended sediments and nutrient-stimulated blooms of epiphytes and plankton attenuate light and
reduce light penetration to seagrass canopies. Light thresholds associated with event-based loss of seagrass
have been derived during recent runoff events. PSII herbicides, which are found above guideline levels in the
GBR, reduce photosynthetic efficiency in seagrass.
Little is known about the types and concentrations of contaminants bound to sediment discharged by rivers
into the GBR and the risk that these pose to GBR ecosystems.
An assessment of pesticide risk to the GBR (Chapter 4) considered pesticides within two groupings: (1) the
additive effects of PSII herbicides (diuron, atrazine, hexazinone, ametryn, tebuthiuron and simazine)
normalised via two methods (toxicity equivalent: TEQ and multiple substances potentially affected fraction:
ms-PAF) and compared to consequence values based on expert opinion and (2) the individual risk of other
pesticides compared to their respective trigger value (in a hierarchical framework from the Great Barrier
Reef Marine Park Water Quality Guideline trigger values, the Australian and New Zealand Water Quality
Guidelines and other internationally-derived values). It was concluded that the Mackay Whitsunday and
Burdekin region are considered to be at highest risk from PSII herbicides, followed by the Wet Tropics,
Fitzroy and Burnett Mary regions. However, the risk of only a fraction of pesticides has been assessed, with
only 6 out of the 34 pesticides currently detected included in the assessment, and therefore the effect of
pesticides is most likely to have been underestimated.
Concentrations of a range of pesticides exceed water quality guidelines in many fresh and estuarine
waterbodies downstream of cropping lands. Of the five ecosystems that were considered within the Great
Barrier Reef, it is suggested that the freshwater reaches of rivers and freshwater/coastal wetlands have the
highest risk from pesticides (PSII herbicides and some non-PSII pesticides) followed by the estuarine reaches
of the rivers, the coastal nearshore zone which includes intertidal and subtidal seagrass meadows, the inner
shelf and the mid and outer shelf.
Plume loading maps have been produced for dissolved inorganic nitrogen and total suspended solids for the
period 2007 and 2011 (Chapter 5). These maps provide an indication of the relative influence of these
pollutants in the GBR during flood plume conditions, classified from low to high. The area of coral reef and
seagrass in each class varies considerably between NRM regions, reflecting the differences in the dispersal
of pollutants and the locations of the ecosystems.
Innovative remote sensing methods to map water clarity can also be used as an interpretative tool for
understanding changes in seagrass meadow area and abundance at large spatial scales. This method was
tested in Cleveland Bay, near Townsville in the northern GBR area (Chapter 6). Methods have been
developed to map plume water types in the GBR by using MODIS true colour images reclassified in function
of their dominant colour. Water types entering into the GBR lagoon though river discharges have been
described as primary, secondary and tertiary, each characterised by different concentrations of Coloured
Dissolved Organic Matter (CDOM), total suspended solids (TSS) and chlorophyll. Substantial seagrass loss has
occurred in Cleveland Bay over the period 2008 to 2011 and changes in seagrass area and biomass were
monitored annually for selected meadows; there was a strong correlation between total bay-wide meadow
area and biomass and exposure to primary water (high TSS and CDOM, low light) at an annual time-scale
and also at a 5 year time scale.
Page 5
Analysing the species composition and size class of phytoplankton associated with various stages of plume
development may help provide the missing link between increased nutrient loads, higher nutrient
concentrations, changed water quality conditions and possible changes to food web/primary production in
GBR waters including COTS outbreaks. Thus it is essential that we have greater knowledge on the drivers
and consequences of changes in water quality, and the associated phytoplankton response. Chapter 8
presents current information from both local and international sources on the potential relationship
between phytoplankton communities and COTS populations, including results of water quality and
phytoplankton community analysis from samples collected in Wet Tropics flood plumes between 2010 and
2012.
Part B: Relative risk of degraded water quality to GBR ecosystems
The results of the Relative Risk Index were combined with the outcomes of the Supporting Studies and published
literature to assess the overall relative risk of degraded water quality to GBR ecosystems (Part B of this report).
A tabulated summary of the overall outcomes is presented in Table i and illustrated in Figure i. In this table we
used the individual assessments of each variable used in the ecological risk assessment to highlight where the
variables dominate among the regions. The primary rank for each variable is listed, with an indication of
whether it dominates in terms of coral reef or seagrass area, or both. The Marine Risk Index for coral reefs and
seagrass is then shown for the results of the combined analysis of water quality variables. The regional
anthropogenic loads as a proportion of the total GBR load for TSS, DIN and PSII herbicides are shown to identify
the primary sources of anthropogenic loads delivered to the GBR, in addition to the Loads Index which combines
this information. Additional information includes facts related to the influence of river discharges to the COTS
Initiation Zone and additional pesticide information. The Relative Risk Index which represents the overall results
of the risk assessment (Part A) is shown as the critical underpinning for the overall conclusions. This is then
informed by the management issues and associated land uses which were derived from published evidence and
expert judgement of the assessment team (informed by the preceding columns). The overall ranking of relative
risk was developed by the assessment on the basis of the overall content of the table.
The information has been coupled with the summary of pollutant sources from land uses in each region to
generate the management priorities identified Table ii. However, cost effective solutions for all of these
management issues are not necessarily currently available for all of these priorities. These issues are discussed
further in the Reef Plan Scientific Consensus Statement 2013 (see Thorburn et al. 2013).
Even though the nutrient related variables of Chlorophyll threshold exceedance and DIN plume loading were
ranked highest in the Fitzroy region, there is insufficient knowledge of the sources of DIN in the Fitzroy region to
make recommendations about management priorities to address nutrients in the region. Further knowledge of
the role of particulate nitrogen, which is largely derived from both cropping and grazing lands, and the
processing of this into dissolved inorganic nitrogen is important for making future management
recommendations in the large grazing catchments of the Fitzroy region.
Overall conclusions:
The main finding of the study was that increased loads of suspended sediments, nutrients (nitrogen and
phosphorus) and pesticides all pose a high risk to some parts of the GBR. However, the risk differs between the
individual pollutants, between the source catchments, and with distance from the coast as per the conclusions
below. These findings are represented in the map shown in Figure ii.
Overall, increased concentrations of nitrogen from catchments between the Daintree and Burdekin Rivers
pose the greatest risk to coral reefs. Runoff from these rivers during extreme and early wet seasons is
Page 6
associated with outbreak cycles of the coral-eating COTS on the northern GBR shelf (15 to 17°S) that
subsequently generate secondary outbreaks throughout the central and southern GBR. GBR-wide loss of
coral cover due to COTS is estimated to be 1.4% per year over the last 25 years, and a new outbreak is
underway. It is estimated that COTS have affected >1000 of the ~3000 reefs within the GBR over the last 60
years.
Of equal importance is the risk to seagrass meadows from suspended sediments (TSS) discharged from
rivers in excess of natural erosion rates, especially the fine fractions (clays). Whether carried in flood
plumes, or re-suspended by wave action, suspended particulate matter creates a turbid water column that
reduces the light required by seagrass and corals. High turbidity has been estimated to affect ~200 inshore
reefs and most seagrass areas. The Burdekin and Fitzroy regions present the greatest risk to the GBR from
increased suspended sediment loads.
Loss of seagrass habitat as a result of cyclones, floods and degraded water quality appears to be associated
with higher mortality of dugong and turtles.
The risk to coastal seagrass beds (and freshwater and estuarine wetlands) from the six commonly used PSII
herbicides (pesticides) was assessed as highest in the Mackay Whitsunday and Burdekin regions, followed by
the Wet Tropics, Fitzroy and Burnett Mary regions to. Concentrations of a range of pesticides exceed water
quality guidelines (thresholds) in many fresh and estuarine waterbodies downstream of cropping lands.
However, the risk of only a fraction of pesticides has been assessed. Only 6 out of the 34 pesticides currently
detected included in the assessment, and therefore the effect of pesticides is most likely underestimated.
The ranking of the relative risk of degraded water quality between the NRM regions in the GBR is: Wet
Tropics > Fitzroy > Burdekin > Mackay Whitsunday > Burnett Mary > Cape York. Priority areas for
management of degraded water quality in the GBR are: Wet Tropics for nitrogen management; Mackay
Whitsunday and lower Burdekin for PSII herbicide management; and Burdekin and Fitzroy for suspended
sediment management.
The regional ranking of risk to coral reefs from degraded water quality is: Wet Tropics > Fitzroy > Mackay
Whitsunday > Burdekin> Cape York > Burnett Mary, while that for seagrass beds is: Burdekin > Wet Tropics >
Fitzroy > Mackay Whitsunday > Burnett Mary > Cape York. The combined assessment of water quality
variables in the GBR used the total area of habitat affected in the highest relative risk areas and end-of-
catchment anthropogenic loads of nutrients, sediments and pesticides added to the GBR lagoon.
Importantly in the Mackay Whitsunday Region, 40% of the seagrass area is in the highest relative risk class
compared to less than 10% for all other regions. However the highly valuable seagrass meadows in Hervey
Bay, and the importance to associated dugong and turtle populations in the Burnett Mary Region, were not
included in the ranking analysis.
The risk assessment presented in this report has generated the most comprehensive assessment of the relative
risk of degraded water quality to GBR ecosystems undertaken to date. However, it is important to reiterate that
the rankings between NRM regions are relative, and do not represent absolute differences in the risk to GBR
ecosystems. In this regard, even the lowest ranked regions of the Burnett Mary and Cape York region may pose
a risk to GBR ecosystems, but relative to the other NRM regions they are considered to be of lower risk. This
information can be useful in guiding investment priorities but should not be used in isolation from other
knowledge related to regionally specific priorities supported by additional evidence, socio economic influences
and limitations to the assessment that may have led to uncertainties in the results. The limitations to the
assessment are described in Section 4 of this report, together with suggestions to improve upon the work
presented herein.
Page 7
Figure i. Map summarising the results of the Loads Index, Marine Risk Indexes (coral reefs and seagrass) and the Relative Risk Index. The assessment for coral reefs also includes
the COTS Influence Index associated with the extent of the influence of regional river discharge to the COTS Initiation Zone; the scores are 100 for the Wet Tropics, and 16 for the
Burdekin region. The method for deriving and combining the indexes is provided in Section 2.4.4.
Page 8
Table i. Summary of the outcomes of the overall assessment of the relative risk of water quality in the GBR. Note that the Burnett Mary Region is shaded in grey to represent the
fact that most reefs and seagrass meadows in this region were not included formally in the analysis and thus the validity of the result has high uncertainty.
Region
Dominant variables in
marine assessment
Variables where max
area is in Region
Marine Risk Index
Regional
Anthropogenic Load
as a proportion of
the Total GBR Load
(%)
Loads
Index
Additional Factors
Relative
Risk
Index
Management
Issues
Associated land
uses
Overall
Ranking of
Relative Risk
CR = Coral Reef
SG = Seagrass
Coral
Reef
Seagrass
TSS
DIN
PSII
Herb
Cape York
COTS Initiation Zone (CR)
12
4
3
<1
<1
0
Influence from catchment
runoff is predominantly
from Wet Tropics Rivers
9
The data in this Region are highly
uncertain due to limited validation
of marine datasets.
LOW
Wet Tropics
100
83
9
20
61
100
86% volumetric
contribution to COTS
Initiation Zone
100
Nutrients
Pesticides
Sugarcane
Bananas
VERY HIGH
Burdekin
TSS 2mg/L (SG, CR)
TSS 7mg/L (SG)
TSS Plume loading (CR,
SG)
Chl 0.45µg/L (SG)
DIN Plume loading (SG)
40
100
32
11
13
62
14% volumetric
contribution to COTS
Initiation Zone
High risk from PSII
herbicides to Ramsar listed
freshwater wetlands in the
lower Burdekin
catchments
76
Sediments
Pesticides
Nutrients
Grazing
Sugarcane
(coastal)
HIGH
Mackay
Whitsunday
Pesticide exposure (CR,
SG)
54
37
4
6
12
25
High risk from PSII
herbicides in Sandy Creek
50
Pesticides
Nutrients
Sugarcane
MODERATE
Fitzroy
TSS 7mg/L (CR)
Chl 0.45µg/L (CR)
DIN Plume loading (CR)
86
59
17
5
4
28
Monitored loads of PSII
herbicides were high in
2011 (not reflected in
modelled baseline)
80
Sediments
Pesticides
Nutrients1
Grazing
Cropping
HIGH
Burnett
Mary
All variables rank
relatively low
11
23
4
4
9
20
The Mary River has the
fourth highest total and
anthropogenic TSS load of
all GBR catchments
19
Sediments
Grazing
UNCERTAIN
All variables rank relatively low,
however, there is high uncertainty
in this result given the lack of data
on the full extent and condition of
corals and seagrass (which are
outside the GBRWHA) available for
this assessment.
1There is insufficient knowledge of the sources of DIN in the Fitzroy region to make recommendations about management priorities for these. Further knowledge of the role of
particulate nitrogen, which is largely derived from grazing lands, and the processing of this into DIN is important for making future management recommendations in the large
grazing catchments of the Fitzroy region.
Page 9
Table ii. Summary of management priorities for reducing the relative risk of pollutants to the GBR.
Relative
Priority
Management Priorities
Region
Pollutant Management
Key land uses
Comments
1
Wet Tropics
Fertiliser nitrogen reduction
Sugarcane,
Bananas
Note that these
actions should not be
prioritised at the
exclusion of other
practices that are
already in place to
manage losses of
other pollutants in the
Regions
Burdekin
Erosion management in Burdekin
Grazing
Fitzroy
Erosion management in Fitzroy
Grazing,
Cropping
2
Burdekin
Pesticide reduction in (lower) Burdekin
and Haughton
Sugarcane
Mackay
Whitsunday
Pesticide reduction in all catchments
Sugarcane
Burdekin
Fertiliser nitrogen reduction in (lower)
Burdekin and Haughton
Sugarcane
3
Mackay
Whitsunday
Fertiliser nitrogen reduction
Sugarcane
Burnett
Mary
Erosion management in all catchments
Grazing
Wet Tropics
Pesticide reduction in all catchments
Sugarcane
Fitzroy
Pesticide reduction in all catchments
Grazing,
Cropping
4
Burnett
Mary
Further information is required to inform the
assessment, including data on the full extent and
condition of corals and seagrass (which are outside the
GBRWHA) in the region.
Habitat mapping,
ecological value
assessment and
monitoring of
ecosystem condition
is required
5
Cape York
Further information is required to understand local
influences.
As a relatively low
impacted area,
management efforts
should aim to
maintain the current
values of the region
Page 10
Figure ii. Illustration of the overall outcomes of the assessment of the relative risk of degraded water quality to Great
Barrier Reef coral reefs and seagrass. The map shows the dominant land uses and priority pollutants and results of the
overall relative risk ranking in each NRM region.
Page 11
1 Introduction
Exposure to land-sourced pollution has been identified as an important factor in the world-wide decline in coral
reef condition (Pandolfi et al. 2003; Burke et al. 2011). Different parts of the Great Barrier Reef World Heritage
Area (GBRWHA) are exposed to different degrees of influence from land-sourced pollutants. The degree of
exposure is a function of factors such as distance from the coast, the magnitude of river discharges, the distance
from river mouths, wind and current directions and, the mobility of different pollutant types, and of course the
different land-uses in the 6 natural resource management (NRM) regions of the Great Barrier Reef (GBR)
catchment. This differential exposure to land-sourced pollutants has important consequences for the likely
degree of degradation that habitats such as coral reefs and seagrass meadows may suffer as a result of land-
sourced pollution and an assessment of exposure is required to prioritise management of such pollution on a
regional basis.
To date the prioritisation of potential management responses between different pollutants, different land uses
and industries, and different NRM regions has used methods such as Multiple Criteria Analysis (MCA) (Brodie
and Waterhouse, 2009; Brodie et al. 2009; Cotsell et al. 2009; Greiner et al. 2005; Waterhouse et al. 2012). The
methodologies used to date, though the best available at the time, were limited for reasons summarised in
Table 1.1. While these analyses have proved useful for ongoing prioritisation of investment under Reef Plan,
more specifically Reef Rescue and the selection of priority management areas under the Reef Protection
Package, more sophisticated analyses are now needed to more confidently prioritise between pollutants, across
the individual catchments. In addition it is worth noting that the availability of data related to seagrass meadows
for these assessments remains relatively sparse.
The assessment approach described in this report is listed in the final column of Table 1.1 below. In 2011, the
Australian Government funded a scoping project through the National Environment Research Program (NERP
Project 4.3: Ecological risk assessment of pesticides, nutrients and sediments on water quality and ecosystem
health) to review the methodology used in previous ecological risk assessment approaches for water quality in
the GBR and make recommendations for a revised and improved assessment approach. The project team
outlined a tiered risk assessment methodology that was recommended to provide a systematic, objective and
transparent approach to quantify the relative risk of pesticides, nutrients and sediment to the GBR (Hayes et al.
2012). A meta-database was compiled to determine the availability of data for the risk assessment. However,
due to timing and data and resource limitations, the proposed method was not able to be adopted by the
Queensland Department of Environment and Heritage Protection (DEHP). Subsequently in 2012, the DEHP
agreed to fund a second phase to the project through the Reef Protection Package Science Program
(http://www.reefwisefarming.qld.gov.au/information/science.html) that required delivery of an assessment of
the relative risk of pollutants to the GBR by March 2013. The approach that was selected for this project builds
on previous assessments (Table 1.1). It still uses an MCA approach; however is improved with new input data,
revised criteria and the application of a spatial multi criteria analysis tool, Multi-Criteria Analysis Shell for Spatial
Decision Support (MCAS-S) developed by ABARES (refer to http://www.daff.gov.au/abares/data/mcass).
Page 12
Table 1.1. Summary of the elements included in past and current assessments of the risk of water quality to the GBR
used to inform Reef Plan management prioritisation.
Element
Reef Plan
Greiner et al. 2005
Reef Rescue
Cotsell et al. 2009
Reef Protection Package
Waterhouse et al. 2012
Reef Plan 3 /
Reef Rescue 2
This report
Method
MCA
MCA
MCA
MCA plus interpretive
studies
Data availability
Very limited
Limited
Moderate
Good
Analysis end point
Coral reefs and
seagrass
Coral reefs
Coral reefs, seagrass,
water column
Coral reefs, seagrass, plus
freshwater to marine
ecosystems for pesticides
Relative
importance of
pollutants
No
No
No
Yes
Pesticide data
Very limited
Limited
Limited
Yes, with limitations
Marine exposure
estimate
Limited – from Devlin
et al. 2003
Moderate – from
Maughan and
Brodie, 2009
Moderate – from
Maughan and Brodie,
2009
Good – from recent work
by Devlin et al. 2013a;
Alvarez Romero et al. 2013
Socio and
economic values
included
Yes
Yes
No
No
Spatial coverage
All GBR, however
Burnett Mary marine
area outside of
GBRMP excluded
All GBR, however
Burnett Mary
marine area
outside of GBRMP
excluded
Cape York and Burnett
Mary excluded
All GBR, however Burnett
Mary marine area outside
of GBRMP excluded
The objective of the project was to estimate the relative risk of pollutants in the GBR catchments to GBR
ecosystem health. The assessment used a combination of qualitative and semi-quantitative information about
the influence of individual catchments in the 6 NRM regions (Figure 1.1) on key GBR ecosystems (coral reefs and
seagrass meadows). The marine boundaries used for each NRM region are those accepted officially by the Great
Barrier Reef Marine Park Authority. The area of GBR lagoon waters within each marine NRM region were also
reported in recognition of other important habitats and populations that exist in these areas.. Qualitative
conclusions were drawn about coastal and estuarine wetlands where information was available. Unlike the risk
assessment undertaken for Reef Plan in 2004 (Greiner et al. 2005), this assessment does not take into account
the social and economic value of the assets such as tourism and fishing values.
This report is presented in 2 main parts:
Part A: Assessment of the relative risk of degraded water quality to ecosystems of the GBR.
Part B: Overall conclusions of the differential risk between sediments, nutrients and pesticides and between
land uses, industries and catchments/regions in the GBR.
Page 13
Supporting Studies (Waterhouse, 2013) were also a key component of this risk assessment and provided
supporting evidence to determine differential risk between sediments, nutrients and pesticides and between
land uses, industries and catchments.
The actual assessment of the relative risk of pollutants to the GBR ecosystems (Part A) is supported by a number
of these specific studies that have either developed or confirmed our understanding of the characteristics of key
pollutants that influence the risk to the GBR ecosystems (Supporting Studies). The results of these studies are
presented as individual chapters in the Supporting Studies report and include:
1. Linkages between river runoff, phytoplankton blooms and primary outbreaks of crown-of-thorns starfish
in the Northern Great Barrier Reef.
2. The Redfield Ratio and potential nutrient limitation of phytoplankton in the Great Barrier Reef.
3. Review of increased suspended sediment delivery to the Great Barrier Reef and the effects of
subsequent sedimentation and light reduction on coral reefs.
4. Assessing the risk of additive pesticide exposure in Great Barrier Reef ecosystems.
5. Mapping of exposure to flood plumes, water types and exposure to pollutants (DIN, TSS) in the Great
Barrier Reef: toward the production of operational risk maps for the World’s most iconic marine
ecosystem.
6. Assessment of seagrass and water quality influences in the Great Barrier Reef: A case study linking
annual measurements of seagrass change to satellite water clarity data (Cleveland Bay, Queensland).
7. Review of the risks to seagrasses of the Great Barrier Reef caused by declining water quality.
8. Analysis of phytoplankton community structure in flood events in the Great Barrier Reef.
Together, the results of these assessments allow conclusions to be drawn about the relative risk of pollutants to
the GBR and relate these back to land uses and catchments (Part B, Section 3).
All of these Chapters and this overall report have been reviewed by independent experts, with the exception of
the report on phytoplankton in Chapter 8 of the Supporting Studies. Only a small amount funding was provided
under this project to undertake the actual analysis of water quality and phytoplankton populations in flood
conditions collected between 2010 and 2012. As limited funding was provided to report on the outcomes of
these analyses, the report of this work provided in Chapter 8 is preliminary and has not been peer reviewed. It is
intended that these results will be considered in the context of a broader review funded through NERP which is
also related to knowledge of GBR phytoplankton populations and the links to nutrient dynamics and COTS
outbreaks. These results will be published at a later date.
The information generated from this assessment will be invaluable to the review of the Department of
Environment and Heritage Protection Reef Water Quality Program (formerly Reef Protection Package) and the
development of Reef Plan 2013 (Reef Plan 3). In addition, it provides an integral component of the overall
investment prioritisation being undertaken for Reef Plan 3 (incorporating Reef Rescue 2) illustrated in Figure 1.2.
The investment prioritisation will take the assessment from this project in association with more detailed
catchment assessments of the analysis of sub catchment pollutant loads to identify catchment hot spots for
pollutant sources, and an assessment of the ‘solvability’ of management issues in different NRM regions and
Page 14
catchment areas. The solvability assessment will be based on recent management practice adoption data for the
sugar cane and grazing industries in the GBR catchments, and the cost effectiveness of improvements in water
quality of the various practices. Together these assessments will provide robust, targeted scientific evidence of
sources of high risk reef pollutants in the GBR catchment, with information to guide the best return on
investment. This will enable better prioritisation of Reef Plan, Reef Water Quality Program and Reef Rescue 2
actions.
The outcomes of this project also inform one chapter of the Reef Plan Scientific Consensus Statement (SCS)
2013; Chapter 3 Relative risks to the GBR from degraded water quality (Brodie et al., 2013b) The SCS chapters
and scope are illustrated in Figure 1.3. The evidence in these chapters also informs the methods and
interpretation of results for this assessment and is referred to throughout this report.
Page 15
Figure 1.1. Map showing the assessment boundaries considered in this risk assessment.
Page 16
Figure 1.2. Assessment framework for the overall Reef Plan 3 and Reef Rescue 2 management prioritisation process. The
Management practice ‘solvability’ assessment will be delivered by the Australian Government for Reef Rescue 2 and
includes analysis of the management practice adoption and cost effectiveness data developed through the Reef Plan
Paddock to Reef Program. The Sources of pollutants assessment is being conducted as a collaborative effort between the
Australian Government and the Queensland Government catchment modelling team. This project delivers the Relative
risk of degraded water quality to the GBR assessment described in this report.
Page 17
Figure 1.3. Scope of the Reef Plan Scientific Consensus Statement 2013 in relation to this risk assessment project.
Page 18
Part A: Assessment of the relative risk of degraded water quality to ecosystems of the GBR
2 Assessment of the risk of pollutants to ecosystems of the Great Barrier Reef including
differential risk between sediments, nutrients and pesticides, and among NRM regions
Jane Waterhouse, Jeffrey Maynard, Jon Brodie, Daniel Zeh, Lucy Randall, Stephen Lewis, Caroline Petus, Michelle
Devlin, Eduardo da Silva, Miles Furnas, Britta Schaffelke, Katharina Fabricius, Vittorio Brando, Len McKenzie,
Catherine Collier, Michael Warne, Rachael Smith, Nyssa Henry, Hugh Yorkston, Dieter Tracey
2.1 Summary of findings
Using the best available knowledge at the time and risk assessment methods that are appropriate within the
resources of this study, we have investigated three key aspects of the relative risk of water quality to the GBR:
(a) the relative importance of different pollutants to GBR ecosystems (coral reefs and seagrass meadows), (b)
the combined risk of degraded water quality on GBR ecosystems, and (c) the relative risk of degraded water
quality in the GBR through consideration of the combined risk and influence of end-of-catchment anthropogenic
pollutant loads (ie. the current load less the estimated pre-European load) on GBR ecosystems.
The risk assessment method used a combination of qualitative and semi-quantitative information about the
influence of individual catchments in the 6 natural resource management (NRM) regions on coral reefs and
seagrass ecosystems. The area of GBR lagoon waters within each marine NRM region were also reported in
recognition of other important habitats and populations that exist in these areas.
A suite of water quality variables were chosen that represent the pollutants of greatest concern with regards to
agricultural runoff and potential impacts on GBR ecosystems. These include ecologically - relevant thresholds for
concentrations of total suspended solids (TSS) and chlorophyll a (Chl) from daily remote sensing observations,
and the distribution of key pollutants including TSS, dissolved inorganic nitrogen (DIN) and photosystem II-
inhibiting herbicides (PSII herbicides) in the marine environment during flood conditions (based on end-of-
catchment loads and surface water exposure estimates). A spatial variable was included that represents the area
of the GBR lagoon where primary crown-of-thorns starfish (COTS) outbreaks have been observed. COTS
outbreaks are an important cause of coral loss on the GBR and appear to be a response to excess nutrient runoff
from certain catchments that impact this ‘COTS initiation zone’. Anthropogenic end of catchment loads were
included to link the outcomes of the marine assessment to catchment management. In recognition of the
importance of the influence of catchment discharges in driving COTS outbreaks, an index of regional
contributions of river discharges to the COTS initiation zone is also included for coral reefs.
For each variable, thresholds above which potential impacts have been observed were defined and classified
into three to five classes (from lowest to highest), largely on the basis of the time (or probability) the ecosystem
is likely to be exposed to concentrations above the threshold; these are defined as ‘assessment classes’.
Key findings:
Relative importance of different pollutants on GBR ecosystems
The area of coral reefs at highest risk from all of the sediment and nutrient variables (except for the
COTS initiation zone) was greatest in the Burdekin and Fitzroy regions. The other regions (Cape York,
Wet Tropics, Mackay Whitsunday and Burnett Mary) each had ~20% or less of the coral reef area
affected for each variable.
Page 19
The area of seagrass at highest risk from the sediment and nutrient-related variables was greatest in the
Burdekin region. The area of seagrass within the Wet Tropics region is second greatest for all sediment-
related variables but the areas are less than one quarter of the areas affected in the Burdekin region in
all cases. For the nutrient variables, the second greatest areas of seagrass within the highest assessment
classes are in the Fitzroy.
The COTS Initiation Zone straddles the boundary between the Cape York and Wet Tropics regions, with
approximately 60% of reefs within the Zone located in the Cape York region.
Of the NRM regions examined in the assessment, the Mackay Whitsunday region presents the highest
ecological risk from pesticides with the Photosystem II (PSII) herbicide risk of ‘High’ and ‘Medium’
extending off the mouths of the Pioneer and O’Connell Rivers and Sandy Creek. This is followed by the
Burdekin (due to the Barratta Creek and Haughton Rivers but not the Burdekin River itself), Wet Tropics,
Fitzroy and Burnett Mary NRM regions. It should be noted that the risk to ‘pesticides’ here is
represented by PSII herbicides as these are the dominant pesticides detected in catchments, however a
total of 34 pesticides (herbicides, insecticides and fungicides) have been detected. In addition the high
risks of PSII herbicides to wetland, estuarine and coastal habitats (which provide important ecosystem
services to the GBR including fish nursery habitats), were not included in this stage of the assessment,
but they are recognised as important and addressed in Supporting Studies Chapter 4 (Lewis et al. 2013a)
and Part B of this report.
Combined risk of degraded water quality to GBR ecosystems
When all water quality variables are combined, the risk is greatest for coral reefs in the Fitzroy and
Mackay Whitsunday regions, and for seagrass in the Burdekin and Fitzroy regions. In most cases, the
proportion of the habitat area in each region that is in the highest risk areas are less than 10%, except in
the case of seagrass meadows in the Mackay Whitsunday region where 37% of the area of seagrass in
the region is affected. This may have significant implications at a regional scale and warrants further
consideration. These high risk areas often include highly valued tourism and recreation sites of the GBR.
Examples include Fitzroy Island, Hinchinbrook Island, Magnetic Island, many of the islands in the
Whitsunday Group and the Keppel Island group. Inshore seagrass meadows are also of critical
importance to dugong and green turtle populations in the GBR.
Relative risk of degraded water quality to GBR ecosystems
A combined assessment of anthropogenic end of catchment pollutant loads and the ecological risk of
water quality variables in the marine environment allows us to draw conclusions about the overall risk
of pollutants to the GBR. In summary, the greatest risk to each habitat in terms of the potential water
quality impact from all of the assessment variables in the GBR and end -of -catchment anthropogenic
loads of DIN, TSS and PSII herbicides is:
- Coral reefs: Wet Tropics region, followed by the Fitzroy region. The rank of the remaining regions is
the Mackay Whitsunday, Burdekin, Cape York and Burnett Mary region.
- Seagrass meadows: Burdekin region, followed by the Wet Tropics. The rank of the remaining
regions is the Fitzroy, Mackay Whitsunday, Burnett Mary and Cape York region.
- Coral reefs and seagrass meadows combined: Wet Tropics, followed by the Fitzroy (80% of this
influence) and Burdekin (76%). The relative risk to the Mackay Whitsunday region is half the
greatest risk, followed by the Burnett Mary and Cape York regions. It must be reiterated that these
Page 20
are relative assessments and therefore indicates that the Burnett Mary and Cape York are low
relative to the other regions, but may be exposed to a range of risks that still warrant management
to maintain the ecosystem values in these regions.
The risk assessment presented in this Section has generated the most comprehensive assessment of the relative
risk of degraded water quality to GBR ecosystems undertaken to date. It is important to reiterate that the
rankings between NRM regions are relative, and do not represent absolute differences. In this regard, even the
lowest ranked regions of the Burnett Mary and Cape York region may pose a risk to GBR ecosystems, but relative
to the other NRM regions they are considered to be of lower risk. This information can be useful in guiding
investment priorities but should not be used in isolation from other knowledge related to regionally specific
priorities, socio economic influences and limitations to the assessment that may have led to uncertainties in the
results. The limitations to the assessment are described in Section 4 of this report, together with ways to
improve upon the work presented herein.
2.2 Introduction
This section of the report presents the results of the most recent effort to assess the relative risk of the
influence of sediments, nutrients and pesticides on key GBR ecosystems. As outlined in Section 1, the
assessment considers the most relevant influences of water quality in the GBR - sediments, nutrients and
pesticides.
This report refers to suspended (fine) sediments and nutrients (nitrogen, phosphorus) as ‘pollutants’. Within this
report we explicitly mean enhanced concentrations of or exposures to these pollutants, which are derived from
(directly or indirectly) human activities in the GBR ecosystem or adjoining systems (e.g. river catchments).
Suspended sediments and nutrients naturally occur in the environment; indeed, all living things in ecosystems of
the GBR require nutrients, and many have evolved to live in or on sediment. The natural concentrations of these
materials in GBR waters and inflowing rivers can vary, at least episodically, over considerable ranges. Pesticides
do not naturally occur in the environment. Pollution occurs when human activities raise ambient levels of these
materials (time averages, or event-related) to concentrations that cause environmental harm and changes to the
physical structure, biological communities and biological functions of the ecosystem.
2.2.1 Risk assessment framework
Ecological Risk Assessment (ERA) is a term used for a variety of methods to determine the risk posed by a
stressor, for example a pollutant, to the health of an ecosystem. “Risk” is usually defined as the probability that
an adverse effect will occur as a result of ecosystem exposure to a certain concentration of the stressor. Risk is
often quantified as the product of the likelihood of an event occurring (exposure) and the consequences (also
measured as effects) of that event. Risk assessments are used as decision tools that rank risks to human values
in order to prioritise management actions and investments (eg. Burgman, 2005; AS/NZS, 2004). A number of
methodologies are available to carry out the analysis with Bayesian techniques now often favoured by decision
makers (e.g. Hart et al. 2005; Hart and Pollino, 2008).
In an initial stage of the current project, an approach using a tiered risk assessment methodology that would
provide a systematic, objective and transparent approach to quantify the relative risk of contaminants to the
GBR ecosystems was developed (Hayes et al. 2012). This tiered ERA approach would have allowed a ranking of
the rivers draining into in the GBR based on exceedance of water quality guidelines at marine sites whose water
quality could be confidently attributed to individual rivers, and across the GBR lagoon as whole. However, this
approach was not able to be used due to limitations in data availability and limitations with time and resources.
Hence a simpler methodology suitable for the existing datasets, resources and timeframes has been developed
based on a modification of the typical ERA framework.
Page 21
In this assessment the relative risk of degraded water quality among NRM regions was determined by combining
information on the estimated ecological risk of water quality to coral reefs and seagrass meadows in the GBR
and end-of-catchment pollutant loads. This approach attempts to relate the water quality conditions in the GBR
to catchment based influences, albeit in a relatively crude way.
Ecological risk is assessed using a relatively simple approach. The likelihood of exposure of a species or habitat to
an impact is typically a function of the intensity of the impact (the concentration or load of a pollutant) and the
length of time it is exposed to the impact. For example, a seagrass meadow may be exposed to a high intensity
impact for a short period of time (acute), or to lower intensities for longer periods (chronic). When quantifying
exposure, it is important to determine the threshold concentrations that lead to an effect on species or habitats,
that is, the concentration that potentially leads to damage or mortality within hours or days, as well as
understanding long-term average concentrations and the duration of exposure. This complicates the description
of exposure thresholds given their values may change by one to two orders of magnitude between days, seasons
and years. Hence, some key water quality variables such as suspended sediments are divided into different
thresholds based on ecological responses and periods of exposure. To reflect this, each threshold is classified
into several assessment classes to represent the potential differences between the duration and severity of the
influence (from lowest to highest).
The consequences are the measured effects of the water quality exposure. Current knowledge of the effects of
degraded water quality on the health of the GBR are summarised in the 2013 Scientific Consensus Statement.
The GBR Water Quality Guidelines reflect our knowledge of ecological thresholds for water quality variables for
coral reefs in the GBR (GBRMPA, 2009). However, only limited information is available to draw conclusions on
the effects of the exposure of sediments, nutrients and pesticides on seagrass health. Evidence shows that one
of the greatest drivers of seagrass health is the availability of light, which is reduced by increased suspended
sediment and the secondary effects of increased nutrients such as increased growth of epiphytes and
phytoplankton (Collier et al. 2012). However, in the absence of more regionally - and species-specific knowledge
of pollutant impacts on seagrass, the same threshold concentrations have been used for coral reefs and seagrass
meadows in this assessment. It is also recognised that the consequence of the exposure of species or habitats to
a range of water quality conditions is complicated by the influence of multiple pressures, and many external
influences including weather conditions, however it is difficult to factor these into the risk assessment in any
quantitative way.
Given the above and recognising the inconsistencies in the spatial and temporal availability of the water quality
data, our capacity to produce a true likelihood or true consequence estimate for this assessment is limited. It
was therefore necessary to develop an effective, simple and standard methodology for the risk assessment that
could be implemented with the available data, in a way that could be easily communicated and discussed with
decision-makers and stakeholders. For this reason, ecological risk in the GBR is expressed simply as the area of
coral reefs and seagrass meadows within a range of assessment classes (very low to very high relative risk) for
several water quality variables in each NRM region in the GBR catchment. Our method for calculating risk
essentially assesses the likelihood of exceedance of a selected threshold. This likelihood was set as 1 for a
parameter and location if observations or modelled data indicate that the threshold was exceeded. Conversely,
the likelihood was set as 0 if observations or modelled data indicate that the threshold was not exceeded. As
consequences are mostly unknown at a regional or species level, potential impact was calculated as the area of
coral reef, seagrass meadows and area of GBR lagoon waters (in km2) within the highest assessment classes of
the water quality variables (reflecting the highest severity of influence). The effects of multiplying the habitat
area by 1 or 0 for the likelihood mean that the final assessment of risk in this assessment is only an indication of
potential impact - the area of coral reef and seagrass meadows in which exceedance of an agreed threshold was
modelled or observed. This becomes an assessment of ‘relative risk’ by comparing the areas of each habitat
Page 22
affected by the highest assessment classes of the variables among NRM regions, and was used to generate a
‘Marine Risk Index’.
Modelled end-of-catchment pollutant loads (generated from the Source Catchments model framework for the
Paddock to Reef Program) were obtained for each NRM region for key pollutants, and only the anthropogenic
portions of total pollutant loads were considered. The anthropogenic load is calculated as the difference
between the long term average annual load, and the estimated pre-European annual load. This information was
used to define a ‘Loads Index’.
The variables included ecologically relevant thresholds for concentrations of total suspended solids (TSS) and
chlorophyll a from daily remote sensing observations, and the distribution of the loading of key pollutants
including TSS, dissolved inorganic nitrogen and photosystem II-inhibiting herbicides (PSII herbicides) in the
marine environment during flood conditions (based on an assessment of flood plume frequency and predicted
distribution of end-of-catchment loads). A spatial variable was included that represents the area of the GBR
lagoon where primary crown-of-thorns starfish (COTS) outbreaks have been observed. COTS outbreaks are an
important cause of coral loss on the GBR and appear to be a response to excess nutrient runoff from certain
catchments that impact this ‘COTS initiation zone’. In recognition of the importance of the influence of
catchment discharges in driving COTS outbreaks, an index of regional contributions of river discharges to the
COTS initiation zone is also included for coral reefs (COTS Influence Index).
The three indexes were then combined to generate a Relative Risk Index for coral reefs and seagrass, which
ultimately ranks the relative risk of degraded water quality to coral reefs and seagrass meadows in the GBR
among NRM regions. The framework and how it relates to the typical ERA framework is shown in Figure 2.1. As
described in the Introduction, the outputs of this assessment provide an integral component of the overall
management and investment prioritisation being undertaken for Reef Plan 3 (incorporating Reef Rescue 2). A
more detailed illustration of the overall process incorporating this assessment is shown in Figure 2.2.
The geographic boundaries of the assessment and the spatial distribution of the marine habitats in the
assessment (coral reefs and seagrass - based on best available information) are shown in Figure 1.1. The area of
GBR lagoon waters in each NRM region was also included as it contains other important habitats and biological
populations such as fish and benthic organisms, however this was not included in the overall Relative Risk Index
as assumptions regarding the importance of the potential impact on the wide range of ecosystems in the GBR
are unknown. The marine boundaries used for each NRM region are those accepted officially by the Great
Barrier Reef Marine Park Authority.
Page 23
Figure 2.1. The risk assessment framework used in this project showing the components of the Marine Risk Index to represent marine water quality
ecological risk to coral reefs and seagrass meadows, a Loads Index to represent catchment influences on GBR water quality using end of catchment
anthropogenic pollutant loads and a COTS Influence Index to factor in the importance of river discharges on the COTS Initiation Zone for coral reefs.
Page 24
Figure 2.2. The overall framework being used for Reef Plan 3 (incorporating Reef Rescue 2) management and investment prioritisation. This project
contributes Steps 1 to 3 and part of Step 4. An additional assessment of the ‘solvability’ of management issues in different NRM regions and catchment
areas is being conducted by the Australian Government. The assessment is based on recent management practice adoption data for the sugar cane and
grazing industries in the GBR catchments, and the cost effectiveness of improvements in water quality of the various practices.
Page 25
2.3 Methods
A 3-part approach of estimating the relative risk of pollutants to the GBR at a regional level was applied
(illustrated in Figure 2.3):
1. Assessment of the relative importance of different pollutants on GBR ecosystems (coral reefs and
seagrass). This identifies the areas where each water quality variable is considered to pose the greatest
relative risk to coral reefs and seagrass between the NRM regions. The output can be used to guide
priorities for management of individual pollutants between NRM regions. The methods are described in
Section 2.3.3.
2. Combined risk of degraded water quality to GBR ecosystems. The combined assessment takes into
account all assessment classes for each variable to generate a Marine Risk Index for coral reefs and
seagrass. The areas within the Risk Index represent the areas of highest relative risk to degraded water
quality in the GBR and identify the areas where coral reefs and seagrass are most likely to be under
pressure from degraded water quality. The methods are described in Section 2.3.4.
3. The relative risk of degraded water quality to GBR ecosystems. This relates the results of Part 1 and Part
2 to land based influences using an assessment of end-of-catchment anthropogenic loads and river
discharges (Loads Index and COTS Influence Index). These results inform the regional management
priorities required to address the risks identified in Part 1 and Part 2 in terms of where to focus effort on
which pollutants. The methods are described in Section 2.3.5.
Justification for the selection and classification of variables is provided in Section 2.3.1. Section 2.3.2 describes
the habitat areas in each region.
2.3.1 Selecting and classifying variables
We have chosen a suite of water quality variables that represent the pollutants of greatest concern with regards
to agricultural runoff and potential impacts on GBR ecosystems. Ecological impacts of terrestrial runoff on coral
reefs and seagrasses beds can be experienced as either acute, short term changes associated with formation of
high-nutrient, high-sediment, low salinity flood plumes or the more chronic impacts associated with changes in
long-term water quality concentration (Devlin et al. 2012). The ecological impact of terrestrial pollutants varies
not only with the type of pollutant, the magnitude and extent of the riverine influence but also with the
ecosystems being affected and the frequency and duration of plume occurrence (e.g. Devlin et al. 2013a). Long
time series of pollutant concentration data provides a way of assessing chronic stress, while river plume models
can help to develop risk maps by defining areas which may experience acute or chronic high exposure to
pollutants or stressors (Alvarez-Romero et al. 2013). Details of the pollutant movement and frequency of
inundation can be key measurements in attributing water quality decline to ecosystem change. This assessment
uses a combination of variables that represent chronic and acute stress.
Page 26
Figure 2.3. Illustration of the steps in the assessment of the relative risk of water quality pollutants to GBR ecosystems.
Page 27
The selected variables are summarised in Table 2.1. These include ecologically relevant thresholds for
concentrations of total suspended solids (TSS) and chlorophyll a from daily remote sensing observations, and the
distribution of key pollutants including TSS, dissolved inorganic nitrogen (DIN) and photosystem II-inhibiting
herbicides (PSII herbicides) in the marine environment during flood conditions (based on end-of-catchment
loads and surface water exposure estimates). A spatial variable is included that represents an area of the GBR
lagoon where primary crown-of-thorns starfish (COTS) outbreaks have most frequently been observed (see
Chapter 1 of the Supporting Studies, Furnas et al., 2013a). COTS outbreaks are an important cause of coral loss
on the midshelf and outer reefs of the GBR (De’ath et al. 2012) and are, based on current understanding, a
response to excess nutrient runoff from certain catchments that reaches this ‘COTS initiation zone’ (Fabricius et
al. 2010). The relevance of each of these variables is described below. More detailed information on pollutant
impacts GBR ecosystems is provided in the recently completed Scientific Consensus Statement Chapter 1 Marine
and coastal ecosystem impacts from degraded water quality (Schaffelke et al. 2013).
For each variable, thresholds above which impacts have been observed or predicted were defined and classified
into three to five classes (from lowest to highest). The classification of each variable is described in Section
2.3.1a, 2.3.1b and 2.3.1c below. The selected variables and thresholds represent long-term conditions (chronic
exposure) and wet season pollutant loadings in flood plumes (acute exposure).
Additional variables were considered that have not been included here due to the current lack of data showing
their temporal and spatial patterns and ecological impacts. These include: phosphorus exposure, chronic
exposure to PSII herbicides and non-PSII herbicides, and time series of pesticide concentration data. The
decision to select DIN as primary nutrient variable within the assessment is supported by the conclusions of
Chapter 2 of the Supporting Studies (Furnas et al. 2013b) which considered the relative importance of nutrient
forms and of nitrogen and phosphorus in the GBR. The analysis indicates that dissolved inorganic and particulate
forms of nutrients discharged into the GBR are both important in driving ecological effects but increased
nitrogen inputs are more important than phosphorus inputs. Dissolved inorganic forms of nitrogen and
phosphorus are considered to be of greatest concern compared to dissolved organic and particulate forms of
nutrients, as they are immediately and completely bioavailable for algal growth (see Furnas et al. 2013b).
Particulate forms mostly become bioavailable over longer time frames, and dissolved organic forms typically
have limited and delayed bioavailability (see Furnas et al. 2013b).
For each of the variables shown in Table 2.1 a classified spatial data layer was prepared. ArcGIS was the primary
tool used for spatial analysis, however, the Multi-Criteria Analysis Shell for Spatial Decision Support (MCAS-S
1
)
was also used for some aspects of the assessment. MCAS-S was initially the preferred tool as it was also used for
the associated Reef Rescue 2 Investment Prioritisation process conducted by the Australian Government.
However, MCAS-S is a raster-based tool that therefore loses a degree of resolution of some spatial layers
depending on the selected grid size. Using a 1 km2 grid, this was problematic for the assessment of the area of
coral reefs because the area was considerably different after the raster conversion. A number of approaches
were tested to overcome these limitations including a presence/absence approach for coral reefs within a grid
cell, but this resulted in significant overestimation of the areas especially in the inshore areas where reef sizes
are typically small or there are many fringing reefs. The differences between regions were not comparable and
therefore, significantly distorted the results of the assessment. Therefore ArcGIS was used for all area estimates
(in km2) for all variables as it is possible to calculate the area within the reef polygon; seagrass habitat is already
on a 1 km2 grid.
1
MCAS-S: Multi-Criteria Analysis Shell for Spatial Decision Support developed by ABARES, refer to
http://www.daff.gov.au/abares/data/mcass
Page 28
MCAS-S is a useful tool for spatial analysis of datasets that has adopted much of the functionality of the ArcGIS
interface but that allows independent data processing and is relatively easy to use. MCAS-S functionality
promotes participatory processes that require the understanding of relationships between decision- making
requirements and the available data. Of particular value in this regard are software features that enable
interactive cognitive and ‘live up-date’ mapping of alternative views (Lesslie et al. 2008).
As described in Lesslie et al. (2008), MCAS-S requires the user to prepare raw spatial data for a project in a raster
format of consistent geographical extent, pixel resolution and projection. Primary data can then be selected
from the menu and dragged into a display workspace whereupon the spatial data layer can be classified using a
variety of simple classification methods and tools for tailoring input data. With the creation of individual class or
rule layers, the user can apply them in weighted combinations to construct composite layers contributing to
themes of interest. Key functions for combining and comparing spatial layers are represented schematically in
Figure 2.4.
MCAS-S Key Functions
A composite layer can be created as a summary of
several input layers. The weighted contribution of any
selected layer to any composite can be set. The
composite map dynamically updates as the user
changes the weightings on input layers and shows a
hierarchical, cognitive ‘map’ showing the development of
individual layers culminating in a final summary layer
representing a theme. Relationships among themes,
specific views and particular indicators may be examined
using several methods.
Two-way comparison enables the user to create a two-
way comparison map, explore the association among
input classes, and define a colour ramp and value scale
to highlight association of high or low values, or feature a
particular geographical region.
Multi-way comparison is used when the spatial
association of two or more data layers is required. The
multi-way analysis uses the radar plot as the basis for
visualisation.
Figure 2.4. Flow chart for MCAS-S showing stages and functionality. Source: Derived from Lesslie et al. (2008).
Page 29
Table 2.1. Summary of water quality variables, assessment classes and data sources included in the marine risk assessment.
Variables
Assessment Class
Data source/methodology
Very Low
1
Low
2
Medium
3
High
4
Very
High
5
Sediments
Total Suspended
Solids (TSS)
concentration (mg/L)
Frequency of
exceedance %
Based on daily observations of TSS in the period 1 Nov 2002 to 30 April 2012. Data
has been interpolated across reefs (which are masked during image processing)
using Euclidean Allocation in ArcGIS. Classification of frequency of exceedance is
based on the number of valid observations in the full observation period. Method
for extraction described in Brando et al. (2013).
Threshold a: 2 mg/L
<1
1-10
10-20
20-50
50-100
Threshold correlates strongly with declines in ecosystem condition such as
increased macroalgal growth and declining diversity. Average annual threshold for
TSS in the Great Barrier Reef Water Quality Guidelines. Refer Section 2.3.1a(i)
Threshold b: 7mg/L
0
<1
1-10
10-20
20-100
Threshold is equivalent to a turbidity of 5 nephelometric turbidity units (NTU).
Shown to have various ecosystem effects including coral reef stress, declines in
seagrass cover (Collier et al. 2012), fish habitat choice, home range movement and
(above 7.5 nephelometric turbidity units) foraging and predator-prey relationships
(Wenger et al. 2013). Refer Section 2.3.1a(ii).
TSS Plume Loading
(mean 2007-2011)
Category 1
Category
2
Category 3
The frequency and extent of the influence of flood plumes containing differing
concentrations of total suspended solids is used to provide an estimation of the
extent of surface exposure of coral reefs and seagrass during wet season
conditions. Modelled using an assessment of plume frequency from satellite
imagery and monitored end of catchment loads in each wet season (Nov to May)
from 2007 to 2011 (Devlin et al, 2013a). The mean of the five annual maps was
selected as a way of factoring in inter-annual variability in river discharge, although
it is recognised that this period was characterised by several extreme rainfall
events. Refer Section 2.3.1c(v).
Page 30
Variables
Assessment Class
Data source/methodology
Very Low
1
Low
2
Medium
3
High
4
Very
High
5
Nutrients
Chlorophyll
concentration (µg/L)
Frequency of
exceedance %
Assessment classes were based on daily observations of Chlorophyll
concentrations over the period 1 Nov 2002 to 30 April 2012. Data was interpolated
across reefs (which are masked during image processing) using Euclidean
Allocation in ArcGIS. Classification is based on the number of valid observations in
the full observation period. Method for extraction described in Brando et al.
(2013).
0.45 µg/L
<1
1-10
10-20
20-50
50-100
Chlorophyll is an indicator of nutrient enrichment in marine waters. De’ath and
Fabricius (2008) identified 0.45 µg/L as an important ecological threshold for
macroalgal cover, hard coral species richness, octocoral species richness. Annual
average threshold for chlorophyll in the Great Barrier Reef Water Quality
Guidelines. Significant benefits for the ecological status of reefs in the Region are
likely if mean annual chlorophyll concentrations remain below this concentration.
Refer Section 2.3.1b(iii).
Dissolved Inorganic
Nitrogen (DIN) Plume
Loading
(mean 2007-2011)
Category 1
Category
2
Category 3
Elevated DIN is an indicator of nutrient enrichment. High concentrations of DIN can
reduce coral recruitment (Babcock and Davies 1991; Loya et al. 2004), enhance
coral bleaching susceptibility (Wooldridge and Done, 2009) and change the
relationship between coral and macroalgal abundance (De’ath and Fabricius,
2010). Elevated concentrations can also be deleterious to seagrass by lowering
ambient light levels via the proliferation of local light absorbing algae thereby
reducing the amount of photosynthesis in seagrass, particularly in deeper water
(Collier, 2013).
Modelled using an assessment of plume frequency from satellite imagery and
monitored end of catchment loads in each wet season (Nov to May) from 2007 to
2011 (Devlin et al, 2013a). The mean of the five annual maps was selected as a way
of factoring in inter-annual variability in river discharge, although it is recognised
that this period was characterised by several extreme rainfall events. Refer Section
2.3.1c(v).
COTS Initiation Zone
Out of
the zone
In the
Zone
Shows an area defined to be highest risk in initiating COTS outbreaks, defined as
the area between Latitude 14.5°S and 17°S and described in Furnas et al. (2013a).
Data from this area shows prolonged periods of high Chl concentrations that
exceed 0.8 µg/L, which is important for COTS larval survival. Refer Section
2.3.1b(iv).
Page 31
Variables
Assessment Class
Data source/methodology
Very Low
1
Low
2
Medium
3
High
4
Very
High
5
Pesticides
PSII Herbicide
modelled
concentration (µg/L)
0.025-0.1
0.1-0.5
0.5-2.3
2.3-10
>10
Based on an estimate of the relationship between Colour Dissolved Organic Matter
(CDOM) and salinity, and then a modelled salinity to PSII herbicide concentration
relationship in a flood plume event in one river in each NRM region in 2009-2011.
Data has been interpolated across reefs (which are masked during image
processing) using Euclidean Allocation in ArcGIS. Risk posed was determined using
a number of methods - some only assessed acute toxic effects, others both acute
and chronic. Described in Lewis et al. (2013a). Refer Section 2.3.1b(iv).
>0.025-0.1 µg/L: No observable effect; 0.1-0.5 µg/L: Photosynthesis is reduced by
up to 10% in corals (Negri et al. 2011); seagrass (Haynes et al. 2000; Chesworth et
al. 2004; Gao et al. 2011; Flores et al. in review) and microalgae (Magnusson et al.
2008, 2010). The effect on primary production is minor. 0.5-2.3 µg/L:
Photosynthesis is reduced by between 10% and 50% in corals (Negri et al. 2011);
seagrass (Haynes et al. 2000; Chesworth et al. 2004; Gao et al. 2011; Flores et al. in
review) and microalgae (Magnusson et al. 2008, 2010). The community structure
of tropical microalgae can be affected by concentrations of diuron as low as 1.6
µg/L (Magnusson et al. 2012). The effect on primary production is moderate. 2.3-
10 µg/L Photosynthesis is reduced by between 50% and 90% in corals (Jones and
Kerswell, 2003; Negri et al. 2011); seagrass (Chesworth et al. 2004; Gao et al. 2011;
Flores et al. in review) and microalgae (Magnusson et al. 2008, 2010). A 50%
reduction of growth and biomass of tropical microalgae was also reported in this
concentration range (Magnusson et al. 2008). The community structure of tropical
microalgae is significantly affected and this causes significant changes in the
tolerance of microbial communities to herbicides (Magnusson et al. 2012). The
effect on primary production is major. > 10 µg/L: reduced growth and mortality in
seagrass (Gao et al. 2011) and loss of symbionts (bleaching) in corals (Jones et al.
2003; Negri et al. 2005).
Page 32
a) Exceedance of suspended solids concentration thresholds
The effects of elevated concentrations of suspended solids on GBR ecosystems including coral, seagrass and
algal communities are described in Chapter 3 of the Supporting Studies (Brodie et al. 2013a). The greatest
influence of increased turbidity caused by resuspension of sediment in waters of depths less than 12 metres is
reduced light for benthic phototropic communities including coral reefs and seagrass (Larcombe et al. 1995;
Anthony et al. 2004; Orpin et al. 2004; Alongi and McKinnon, 2005). This resuspension driven turbidity persists
for many months of the year in GBR coastal waters. Suspended solids in flood plumes also reduce light for
benthic communities but the effects are only present for short periods, typically days to weeks. Hence, a long -
term time series is most relevant in the assessment of chronic effects of elevated suspended solids and turbidity
on habitats. However, typically the resuspended sediment is that which was delivered as a sediment loading
during the previous wet season and potentially earlier wet seasons as well. Hence, there is a strong connection
between turbidity and river loadings of sediment (Fabricius et al. 2013). The TSS plume loading modelling (Devlin
et al. 2013a) allows us to assess loadings and in a sense predict the likely conditions of suspended sediment in
various areas of the GBR lagoon. When fine sediment is delivered to shallow areas less than 12 metres, it is a
good indicator for likely resuspension later in the year. Therefore, both the concentration data and the loading
data are relevant to this assessment. The actual exposure of benthic organisms (for example in Cleveland Bay) to
flood plume turbidity is more relevant for the assessment of acute effects (see Devlin et al. 2013b).
Method:
Using remote sensing imagery in the period 2002 to 2012, we defined the areas where the TSS concentration
exceeded of the different ecologically-relevant threshold values at different frequency intervals (see Table 2.2).
The method for the retrieval and processing of the remote sensing data is described in Brando et al. (2013) and
is summarised below.
Data collected by MODIS Aqua provide a time series from 1 November 2002 to 30 April 2012 of water quality
estimates with spatial coverage at 1 km2 resolution for the whole-of-GBR lagoon, nominally on a daily basis
(except overcast days). The water quality estimates were retrieved from the MODIS Aqua time series using two
coupled physics-based inversion algorithms developed to accurately retrieve water quality parameters for the
optically complex waters of the GBR lagoon (Brando et al. 2008; Schroeder et al. 2008; Brando et al. 2010a,b;
Brando et al. 2012; Schroeder et al. 2012). This was necessary because chlorophyll concentrations retrieved with
the MODIS standard algorithms provided by NASA are up to two-fold inaccurate in GBR waters (Qin, Dekker et
al. 2007), while CSIRO’s regionally parameterised algorithms account for the significant variation in
concentrations of Colour Dissolved Organic Matter (CDOM) and TSS and achieve more accurate retrievals
(Brando e al. 2010a,b). For this work the whole MODIS Aqua time series was reprocessed with the most recent
updates in NASA’s software (SeaDAS version 6.4), incorporating the improved knowledge of instrument
temporal calibration to improve temporal stability of the time series of the MODIS Aqua ageing sensor.
The comparison of MODIS Aqua retrievals of Chl, CDOM and turbidity data to in situ data showed that the a-LMI
water quality algorithm coupled with the ANN atmospheric correction is more accurate than NASA’s algorithms
for GBR waters (Brando et al. 2013). The parameterisation and validation on the remote sensing retrievals was
mainly based on observations performed in coastal and lagoonal waters during the dry season between Keppel
Bay and the Wet Tropics region. The accuracy of the retrieval is likely to be lower in shallow and turbid waters
systems such as Princess Charlotte Bay, Broadsound and Shoalwater Bay, as there is no data available for
parameterisation and validation. Details on the algorithm’s theoretical basis, parameterisation and validation
are provided in Brando et al. (2013).
Page 33
The frequency of exceedance of the TSS threshold was calculated by analysing ‘daily’ observations of TSS
concentrations at a scale of 1 km2 pixels. The assessment classes were defined using the total potential number
of observations, and the maximum number of valid observations in the assessment period (Table 2.2). The low
number of valid observations is a result of the strict quality control criteria applied to the imagery: pixels with
cloud or cloud shadow, low view and illumination angles (solar zenith and observer zenith higher than 60
degrees) are flagged and dismissed as are pixels where the atmospheric correction failed. In the wet season a
valid observation is obtained approximately 1 in every 5 days (23%), while for the dry season valid observations
were obtained approximately 2 of every 5 days (41%), equating to less than 2 valid observations every 5 days
over the full year (32%). In a ten-year period, the total potential number of observations is 3,650. The
assessment classeswere defined using expert opinion, informed by Jenks natural breaks on the basis of the
frequency of exceedance of the thresholds (<1%, 10%, 20%, 50% and 100%) as a proportion of the total number
of valid observations.
Table 2.2. Number of valid remote sensing observations (in the context of the total potential observations) throughout
the assessment period, and the frequency of exceedances used to define assessment classes for the assessments. The
selected variables are all Annual and highlighted in grey shading.
Item
Number of observations
Wet
season
10 yrs
Dry Season
10 yrs
Annual
10 yrs
No Pixels (maximum potential observations)
1 pixel = 1 day
1820
1820
3650
Valid observations (pixels with data in assessment period)
427
755
1182
Actual as proportion of potential observations
23%
41%
32%
Frequency of exceedance (based on valid observations)
<1% exceeding thresholds
<4
<7
<10
10% exceeding threshold
43
76
118
20% exceeding threshold
85
151
236
50% exceeding threshold
214
378
591
100% exceeding threshold
427
755
1182
Assessment classes:
Several ecologically-relevant values were explored for TSS using the wet season, dry season and annual datasets
over the 10-year period. The threshold concentrations considered in the assessment are the GBR Water Quality
Guideline value of 2 mg/L and turbidity 5 NTU (which equates to TSS 6.6 mg/L but is rounded to 7 mg/L for
reporting only). The justification for these is described below and further detailed in Chapter 3 the Supporting
Studies (Brodie et al. 2013a).
i Threshold a - TSS 2 mg/L
The assessment classes for TSS were obtained from the annual dataset from the frequency of exceedance
(expressed as a percentage) of the Threshold a concentration of 2 mg/L (1.5 NTU) which is the TSS Water Quality
Guideline value (GBRMPA, 2009). Fabricius (2011) reviewed the factors determining the vulnerability of specific
coral reefs to damage from turbidity and sedimentation. The study found that well-flushed locations with strong
currents, shallow reef crests surrounded by a deep water body, and reefs inhabited by healthy populations of
Page 34
fishes are likely to have the highest levels of resistance and resilience to degradation from sedimentation and
high or variable turbidity. In contrast, damage is most likely to occur in locations with weak currents such as
embayments (e.g. Keppel Bay, Cleveland Bay, Missionary Bay), and within these sheltered zones on deeper reef
slopes, in places where fish abundances are low, and in regions that are frequently affected by other forms of
disturbance such as cyclones , bleaching or crown-of-thorns starfish predation. Despite these complications,
water clarity with an annual mean of less than 10 m Secchi depth (or TSS concentration greater than 2 mg/L),
has been found to relate to strongly altered ecosystem properties throughout the GBR. These values have
therefore been adopted as Water Quality Guidelines for inshore waters by the Great Barrier Reef Marine Park
Authority (GBRMPA, 2009).
As described above, the degree of exposure of an organism or ecosystem to a stressor is typically a function of
the both amount (= concentration, level or load) of the stressor, and the length of time it is in contact with the
stressor (Figure 2.5). A reef may be exposed to a high level of a stressor for a short period of time, or to lower
levels for longer periods. When quantifying exposure levels, it is therefore important to determine peak
concentrations (potentially leading to damage or mortality within hours to days), as well as to quantify long-
term mean (or median) concentrations and the duration of exposure. This complicates the definition of
exposure thresholds in turbidity and sedimentation given their values may change by one to two orders of
magnitude between days, seasons and years. It is for these reasons that we have considered two different
thresholds of TSS in this assessment.
Figure 2.5. Conceptual representation of the exposure of corals to
turbidity and suspended solids concentration, and the severity of
response, are a function of exposure time and the concentration of the
stressor (from Fabricius 2011). The red dotted line shows the GBR WQ
Guideline value for TSS of 2 mg/L.
The data extracted from the remote sensing assessment was classified into the assessment classes shown in
Table 2.1 are based on the total number of daily observations that exceeded the threshold value over the 10-
year assessment period as a proportion of the total number of valid observations (Table 2.2).
ii Threshold b - TSS 7 mg/L (turbidity 5 NTU)
The annual assessment data for TSS 6.6 mg/L (rounded to 7 mg/L for reporting here) was used in the analysis
and was correlated with effects on several ecosystems:
Coral reefs - Cooper et al. (2008) suggest long-term turbidity concentrations >3 NTU lead to sub-lethal
stress, whereas long-term turbidity concentrations >5 NTU correspond to severe stress effects on corals
at shallow depths.
Seagrass – change in seagrass cover at Picnic Bay showed a strong correlation between a decline in
seagrass cover and mean turbidity greater than 4 NTU over several days (Collier et al. 2012; Chapter 6 of
Supporting Studies, Devlin et al. 2013b). However, this effect was observed using a small dataset and
requires further investigation.
Page 35
Fish – Wenger et al. (2013) showed that turbidity of 5 NTU was the threshold for fish habitat choice and
home range movement, and further Wenger et al. (2012) showed that 7.5 NTU was the threshold for
foraging ability and changes in predator prey relationships.
The data extracted from the remote sensing assessment was classified into the ‘frequency of exceedance’
classes shown in Table 2.2. These classes then were used to establish the relative severity classes (very low to
very high) shown in Table 2.1 and are based on the proportion of daily observations that exceeded the threshold
value over the 10-year assessment period. Note that the classes are different to those for threshold 1 and reflect
the increased severity of the impacts associated with this higher TSS concentration threshold.
b) Chlorophyll concentration exceedance
Chlorophyll (Chl) concentrations are relevant year round as an indication of nutrient enrichment in marine
waters. Chl concentrations are particularly relevant when considering drivers of Crown of Thorn Starfish (COTS)
outbreaks. The threshold concentrations for Chl related to adverse effects on coral reefs are well established,
but only preliminary results are available to relate Chl concentrations to seagrass health. Further discussion of
the impacts of nutrient enriched waters on GBR ecosystems is described in the Reef Plan SCS Chapter 1 Marine
and coastal ecosystem impacts (Schaffelke et al. 2013).
Method:
Using remote sensing imagery in the period 2002 to 2012, we defined the areas where the Chl concentration
exceeded of the different ecologically-relevant threshold values at different frequency intervals (see Table 2.2).
The method for the retrieval and processing of the remote sensing data is described in Brando et al. (2013) and
is the same as that applied for TSS exceedance described in Section 2.3.1a above.
Assessment classes:
Several ecologically-relevant values were explored for Chl using the wet season, dry season and annual datasets
over the 10-year period. These ranged between 0.13 µg/L which considered to be insignificant in terms of
potential ecological impacts, various concentrations within the GBR Water Quality Guidelines (seasonal and
cross shelf differences) including 0.45 µg/L, and the highest threshold of 2 µg/L which is known to be associated
with severe ecosystem impacts. Based on expert opinion and data availability the final concentration included in
the assessment was Chl 0.45 µg/L, described below.
iii Chl 0.45 µg/L
Syntheses and interpretation of long-term Chl a and turbidity datasets were presented in De’ath and Fabricius
(2008) to underpin the development of the GBR Water Quality Guidelines. For Chl, an analysis of the response of
macroalgal cover, species richness of hard corals, and species richness of phototrophic and heterotrophic
octocorals, coupled with an assessment of the spatial distribution of water quality, concluded that there would
be significant benefits ecological status of reefs in the GBR if the mean annual Chl concentration was kept below
0.45 µg/L in both coastal and inner shelf zones in all regions.
In addition, a variety of experimental, modelling and observational evidence indicates that initiation of COTS
outbreaks is coupled to enhanced survival of the pelagic larval phase as a result of increased food availability,
particularly if Chl concentrations are greater than 0.45 µg/L.
For these reasons we selected the value of Chl 0.45 µg/L as the most appropriate threshold representing long-
term average conditions.
Page 36
There is less information available on the relationship between seagrass health and chlorophyll concentrations
however it is clear that elevated chlorophyll concentrations increase light attenuation and reduce light
availability to seagrasses (Chapter 7 of Supporting Studies, Collier, 2013). However, preliminary assessments by
Collier et al. (in review) show that decline in seagrass cover is correlated with increasing average daily Chl
concentrations above 0.3 µg/L, however, there is insufficient information to differentiate this response from
those observed at the Guideline value of 0.45 µg/L. This value has been adopted in the GBR Water Quality
Guidelines as also relevant to seagrass, and in the absence of any further information has also been used in this
risk assessment for seagrass.
The data extracted from the remote sensing assessment was classified into the ‘frequency of exceedance’
classes shown in Table 2.2. These classes were used to establish the relative severity classes (very low to very
high) shown in Table 2.1 and are based on the total number of daily observations that exceeded the threshold
value over the 10-year assessment period.
iv COTS specific assessments
Two of the potential threshold concentrations for Chl obtained from the remote sensing data were based on the
relationship between survival of COTS larvae and Chl concentrations (see Fabricius et al. 2010 and Figure 2.6
below). These values were Chl concentration of 0.8 µg/L where COTS larval survival starts to increase, and 2 µg/L
where 100% COTS larval survivorship occurs.
Figure 2.6. Relationship between chlorophyll concentrations and the
probability of development of COTS through larval survival.
Note: mg m-3 is equivalent to µg/L. Source: Fabricius et al. (2010).
The case study analysis presented in Chapter 1 of the Supporting Studies (Furnas et al. 2013a) strongly supports
the hypothesis that river discharge (and associated nutrient loads) is the primary driver of regional chlorophyll
concentrations within the GBR lagoon. The critical factor is the amount of discharge during the early part of the
wet season (November-February) when COTS larvae are present in the water column. There appears to be an
important threshold on the order of 10,000,000 ML (10 km3) of discharge in this period to produce sufficient
phytoplankton at the proper time to sustain high COTS larval survival and promote a subsequent COTS outbreak.
Further development of a COTS outbreak, however, depends on there being sufficient coral cover to sustain
local adult populations.
Based on this information and the importance of the timing of wet season discharges, it would have been
beneficial to undertake further analysis of the long term Chl dataset for the period 1 November to 28 February
each year over the 10 years of available data. However, given the limitation of time and resources associated
with the project it was not able to be completed in time for this assessment.
Page 37
In the absence of this more detailed assessment of Chl concentrations in specific periods, a high risk area for the
initiation of COTS primary outbreaks has been defined. A variety of experimental, modelling and observational
evidence indicates COTS primary outbreaks are initiated by an episode of greatly enhanced larval survival due to
increased food availability (chlorophyll concentrations >0.45 µg/L) for the filter-feeding pelagic larval stage
(Brodie et al. 2005). Observational evidence indicates that primary COTS outbreaks originate on reefs within a
~250 km sector between Cairns and Lizard Island (17.2–14.5S latitudes). Primary COTS outbreaks in the Cairns-
Lizard Island region since 1960 (n=4) have occurred 2-5 years after wet seasons when aggregate early wet
season (November-February) discharge from the Burdekin to Daintree Rivers exceeded 10 km3. The 2-5 year
delay is consistent with the time it takes for COTS to grow up from small (<0.5 cm) recruiting juveniles with
cryptic behaviour at settlement to conspicuously large (>50 cm) and non-cryptic adults of breeding size that
produce large feeding scars. The available data and more recent satellite imagery shows that extensive flooding
from central GBR and Wet Tropics rivers leads both to large nutrient inputs (Furnas, 2003) and widespread
phytoplankton growth (blooms) over large areas of the GBR shelf, including the region between Cairns and
Lizard Island. Regional chlorophyll concentrations in the nominal outbreak initiation region (1989-2012) are
inversely related to regional salinity which is most strongly influenced by wet season rainfall and river runoff
(more rainfall = more runoff = lower salinities). Regional chlorophyll concentrations during the critical pelagic
larval period (Nov-Feb) are positively correlated with estimates of concurrent river runoff at both the local and
regional scales (Furnas et al. 2013a). The ‘high-risk’ area for primary COTS outbreaks between 14.5S and 17S
contains approximately 200 reefs with an aggregate outer reef slope circumference close to 2,000 km (GBRMPA
GIS Group, personal communication).
In this assessment, we have used a binary assessment approach where the reefs are either within or outside of
the COTS Initiation Zone. As shown in Table 2.1, the areas are either classified as Very Low or Very High
depending on whether the reefs are located within the COTS Initiation Zone.
c) Pollutant loading in river plumes
Ecological impacts of terrestrial runoff on coral reefs and seagrass meadows can be experienced as either acute,
short term changes associated with formation of high-nutrient, high-sediment, low salinity flood plumes or the
more chronic impacts associated with long-term changes in water quality (Devlin et al. 2012). The ecological
impact of terrestrial contaminants varies not only with the type of pollutant, the magnitude and extent of the
riverine influence but also with the ecosystems being affected and the frequency and duration of plume
occurrence (see for example Devlin et al. 2012). River plume models can help to develop risk maps by defining
areas which may experience acute or chronic high exposure to pollutants or stressors (Alvarez-Romero et al.
2013). Details of the pollutant movement and frequency of inundation can be key measurements in attributing
water quality decline to ecosystem change. These contribute to the ‘likelihood’ component of the risk equation.
Land sourced runoff containing elevated nutrient concentrations results in flood plumes in the GBR lagoon
which may result in a range of impacts on coral communities (Fabricius et al. 2005; Fabricius, 2011; Brodie et al.
2011). Dissolved inorganic and particulate forms of nutrients discharged into the GBR are both important in
driving ecological effects but increased nitrogen inputs are more important than phosphorus inputs (see Chapter
2 of the Supporting Studies, Furnas et al. 2013b). Dissolved inorganic forms of nitrogen and phosphorus are
considered to be of greatest concern compared to dissolved organic and particulate forms of nutrients, as they
are immediately and completely bioavailable for algal growth. Particulate forms mostly become bioavailable
over longer time frames, and dissolved organic forms typically have limited and delayed bioavailability (Furnas et
al. 2013b).
Most studies in GBR waters show that high levels of dissolved inorganic nitrogen and phosphorus can cause
significant physiological changes in corals, but do not kill or greatly harm individual coral colonies (reviewed in
Page 38
Fabricius, 2005). However, exposure to dissolved inorganic nitrogen can lead to declining calcification, higher
concentrations of photo-pigments (affecting the energy and nutrient transfer between zooxanthellae and host;
Marubini and Davies, 1996), and potentially higher rates of coral diseases (Bruno et al. 2003). Macroalgae and
heterotrophic filter-feeders benefit more from dissolved inorganic and particulate organic nutrients than do
corals. As a result, corals that can grow at extremely low food concentrations may be out-competed by
macroalgae and/or more heterotrophic communities that grow best in high nutrient environments (Fabricius,
2011). Densities of benthic filter feeders – such as sponges, bryozoans, bivalves, barnacles and ascidians –
increase in response to nutrient enrichment (Costa Jr et al. 2000). In high densities some filter feeders, such as
internal macro-bioeroders, can substantially weaken the structure of coral reefs and increase their susceptibility
to storm damage. Critically, more recent research shows that direct interactions between nutrients species such
as nitrate and enhanced coral bleaching susceptibility will be important as a clear example of direct synergy
between climate change stress and nutrient enrichment stress (Wooldridge 2009a; Wooldridge and Done 2009).
The impacts of nutrients on seagrass are less well known and there has been limited, detailed exploration of
nutrient dynamics and nutrient limitation in the GBR, with notable exceptions (Udy et al. 1999; Mellors, 2003).
Therefore, nutrients as an environmental driver has so far been difficult to elucidate because of other over-
riding factors such as light limitation, which tends to be a primary driver (Collier and Waycott 2009). Nutrient
enrichment can stimulate seagrass growth (Udy and Dennison 1997; Udy et al. 1999) if other factors, such as
light availability, are not limiting (Chapter 7 of the Supporting Studies, Collier, 2013). Although a theoretical
nutrient toxicity level does exist, nutrient over-enrichment tends to impact at ecosystem scales and follow a
path of eutrophication with excessive production of organic matter. In addition, nutrients favour the growth of
plankton, macroalgae and epiphytic algae, all of which attenuate light to seagrass leaves (Collier, 2013). In the
GBR some very high epiphyte loads occur on seagrass meadows of the GBR (Mckenzie et al. 2012) and are likely
to reduce light reaching seagrass leaves. However, to date, these have largely been seasonal blooms, and
epiphyte cover has not correlated well with seagrass abundance (Mckenzie et al. 2012). Although nutrient
enrichment has been linked to high algal cover (Campbell et al. 2002), seagrass loss has rarely been attributed to
nutrient over-enrichment. Further discussion of the impact of flood plumes and degraded water quality on
seagrass ecosystems in the GBR is included in Chapter 6 of the Supporting Studies.
The impacts of elevated suspended solids to coral reefs and seagrass meadows were described above in Section
2.3.1a. It is important to note that particulate matter in plumes changes from ‘clay’ (or mineral)-based material
in inshore regions, to organic matter (algal material) in offshore regions. These different types of particulate
matter can have different effects on coral reefs and seagrass meadows as described in Chapter 3 of the
Supporting Studies (Brodie et al. 2013a).
Further discussion of the impacts of TSS and DIN on GBR ecosystems is provided in SCS Chapter 1 (Schaffelke et
al. 2013). Given the importance of flood plumes in delivering TSS and DIN to the GBR, the following variables
related to plume loadings have been included in the assessment.
v DIN and TSS Plume Loading (mean 2007-2011)
Method:
Plume loading maps have been developed for TSS and DIN over the period 2007 to 2011. To date the method
has not been developed for phosphorus due to limitations in our understanding of the relationship between end
of catchment loads of phosphorus and marine concentrations due to complex dynamics in phosphorus
processing in marine waters. However, in the assessment of the relative importance of nitrogen and phosphorus
(Chapter 2 of the Supporting Studies, Furnas et al. 2013b) it was concluded that increased nitrogen inputs are
more important than phosphorus inputs in driving ecological effects in the GBR.
Page 39
The method for assessing plume loadings is described by Alvarez Romero et al. (2013), with further detail in
Chapter 5 of the Supporting Studies (Devlin et al. 2013a). The method involves:
i. Classification of GBR surface waters into colour classes corresponding to plume surface water, non-
plume surface water, cloud and sun glint areas;
ii. Creation of weekly plume colour class maps and mean annual color classes by overlaying maps created
in (i). Weekly composites maps minimize the amount of area without data per image due to masking of
dense cloud cover, common during the wet season (Brodie et al. 2010), and intense sun glint;
iii. Creation of maps of annual frequency of occurrence of plumes, by overlaying weekly composites
created in (ii). These maps help evaluating the annual frequency of occurrence of plumes, representing
the number of weeks plumes are present (i.e. classified as plume surface water) in every pixel/areas of
the GBR and during wet seasons;
iv. Creation of spatially distributed TSS and DIN loading maps. This step involves: (a) the calculation of the
percentage of the TSS and DIN load delivered by each of the 7 GBR priority rivers (Joo et al. 2012; Reef
Plan 2003) in relation to the total TSS and DIN load from the catchments in Wet Tropics, Burdekin,
Mackay Whitsundays and Fitzroy NRMs, and (b) the creation of grids representing the annual average
distribution of TSS and DIN load delivered by the seven major rivers in the study region by multiplying
their proportional contribution to the region-wide TSS and DIN loads with a cost-distance grid (see
Alvarez-Romero et al. 2013) defining the maximum area of influence and the dispersal of pollutants in
the sea. The individual spatially distributed grids (one per river) are then summed to represent the full
TSS load per cell; the overlap of two or more grids defined cells influenced by multiple rivers;
v. The production of annual risk maps for TSS and DIN loading within the GBR by multiplying the annual
frequency of plume occurrence grid (iii above) by the grid representing the sum of spatially distributed
TSS and DIN loads for all rivers (iv above). Exposure values are finally grouped in 5 categories of
exposure (from very low to very high) to investigate spatial variation in exposure; and
vi. Annual exposure maps were then reclassified into three categories of risk (high, medium, low – see
Table 2.1). For this assessment the mean of the 5 annual maps in the period 2007 to 2011 was selected
as a way of factoring in some inter-annual variability, although it is recognised that this period was
characterised by several extreme rainfall events.
The above assessment approach was not completed for Cape York because of outstanding issues with validation
of true colour in this area. However, in recognition that there is some influence of river plumes in the Cape York
region, a more crude estimate of TSS and DIN plume loading in this region has been applied which is
summarised here and described in more detail in Chapter 5 of the Supporting Studies (Devlin et al. 2013a). This
involved:
a) Defining the primary and secondary plume types in the GBR for 2011 (see Alvarez-Romero et al. 2013;
Devlin et al. 2013a).
b) For TSS plume loading: Using the extent of the primary plume type in Cape York to define the extent of
surface exposure. This is where coastal waters are characterised by elevated Colour Dissolved Organic
Matter (CDOM) and TSS, with TSS concentrations dropping out rapidly as the heavier particulate
material flocculates and settles to the sea floor. This area was then allocated as Low exposure as a
conservative estimate given that there is limited information on the frequency of the occurrence of
Page 40
these conditions, and low confidence in the remote sensing outputs in this Region. The remaining area
in the Region was assessed as having no plume exposure.
c) For DIN plume loading: Using the extent of the secondary plume type in Cape York to define the extent
of plume exposure. This is where intermediate waters are characterised by a region where CDOM is
elevated, TSS concentrations are reduced due to sedimentation, and the increased availability of light
and nutrient availability prompt phytoplankton growth measured by elevated Chl concentrations. This
area was then allocated as Low exposure as a conservative estimate given that there is limited
information on the frequency of the occurrence of these conditions, and low confidence in the remote
sensing outputs in this Region. The remaining area in the Region was assessed as having no plume
exposure.
Note that this assessment for Cape York is not included in the overall normalisation of loads for the model
described in Step (iv) above, however, exposure scores for pollutants, particularly DIN, are typically low for the
northern GBR and would not add significantly to the calculation of the normalised loads.
We also assessed the option of just using the plume loading maps from 2011 as a representation of a year of
peak flood conditions, and for comparability to the pesticide exposure maps. However, it was concluded that
the results were too biased to the influence from the large rivers of the Burdekin and Fitzroy which experienced
flow conditions well above long term median records in 2011. Accordingly, the mean of the full set of available
data (2007-2011) was used to represent variation in the characteristics of river discharges and hence flood
plumes between the catchments between years.
Assessment classes:
The following three relative pollutant loading classes were allocated: Low, Medium and High. As an indication of
the relative differences between these classes, preliminary satellite/in-situ match-up analyses have been
performed to validate the 2007-2011 annual exposure maps and relate in situ data to the plume loading
(exposure) classes and described in Chapter 5 of the Supporting Studies (Devlin et al. 2013a). There is a strong
correlation between the in situ data and the assessment classes for DIN and TSS. For example, DIN
concentrations increased linearly (R2 = 0.95) with the highest concentrations corresponding to the highest DIN
exposure category. This reflects the conservative mixing that has been described for DIN in the GBR lagoon
(Devlin and Brodie, 2005; Devlin et al. 2012). TSS concentrations still increased along an increasing gradient of
the level of exposure, with a rapid rise in concentrations in the highest exposure categories (R2 = 0.95). This
exponential increase reflects the sedimentation of the larger, heavier particles in the low salinity zones, and the
transport of the finer sediment over much larger spatial scales. This sediment process is described for the
Burdekin (Bainbridge et al. 2012), Tully (Devlin and Schaffelke, 2009) and reflects the higher risk associated with
the availability of the finer sediment over longer time scales (Fabricius et al. 2012; Brodie et al. 2012b).
d) Pesticide concentrations
vi PSII Herbicide modelled concentration 2009-2011
Waters of the GBR lagoon are contaminated with a range of pesticides including herbicides, insecticides and
fungicides. Pesticides, unlike nutrients, sediments and metals, have no natural sources and their concentrations
have been positively correlated with low salinity associated with river runoff (Lewis et al. 2009; Kennedy et al.
2012a). Therefore, the occurrence of pesticides in the GBR can be attributed with great confidence to
agriculture in the catchments that result in river discharge into the GBR lagoon. Of the 34 pesticides that have
been detected in catchments draining to the GBR, several persistent and mobile PSII herbicides dominate the
Page 41
pesticides identified in water samples and passive samplers in both near-shore and offshore sites on the GBR.
Further information describing the relevance of PSII herbicides to this assessment is described in Chapter 4 of
the Supporting Studies (Lewis et al. 2013a).
Multiple PSII herbicides are usually detected in water samples from the GBR (Lewis et al. 2012) and their
combined effects on microalgae are additive (Shaw et al. 2009; Magnusson et al. 2010). This additive toxicity is
not currently addressed in regulatory guidelines (King et al. in press; Lewis et al. 2012) and is considered to be
important in this assessment. The reduced photosynthesis in algae due to herbicide exposure causes reductions
in the growth of these algae (Magnusson et al. 2008) and changes in species composition (Magnusson et al.
2012) but the effects of chronic exposures in near-shore environments remain largely unknown. This assessment
incorporates an assessment of the acute exposure of PSII herbicides in the 2009-11 wet seasons.
Method:
A full description of this method is provided in Chapter 4 of the Supporting Studies (Lewis et al. 2013a). A
modelling approach based on the relationship between CDOM and sea surface salinity (Schroeder et al. 2012),
was used with the results of in situ end of catchment and GBR lagoon pesticide concentration results for the
2009-2010 and 2010-2011 wet seasons.
Pesticide concentrations were assessed at the end-of-catchment monitoring sites in the 2009-2010 and 2010-
2011 water years (Smith et al. 2012; Turner et al. 2012, 2013) to identify the periods where the higher
concentrations coincided with elevated stream flows (based on the gauges of the Queensland Department of
Natural Resources and Mines; QDNRM, 2012). Moderate Resolution Imaging Spectroradiometer (MODIS) Level-
0 data with 1 km2 resolution were acquired from the NASA Ocean Colour website
(http://oceancolor.gsfc.nasa.gov). The most appropriate satellite image (i.e. the most free of cloud cover and
sun glint) was selected for each NRM region within one week following the highest PS-II concentration. MODIS
images were processed with the SeaWiFS Data Analysis System (SeaDAS). The semi-analytical model developed
by Garvel-Siegel-Maritorena (GSM, Maritorena et al. 2002) implemented in SeaDAS was used to retrieve the
absorption coefficient for dissolved and detrital material (CDOM+D). Bio-optical algorithms often fail to retrieve
correct information over reef bottom type. Pixels values corresponding to reef locations were thus masked out
from the CDOM regional maps. The Cape York region was excluded from this process due to the lack of
monitoring data and the limited use of pesticides in this region.
CDOM was extracted from the satellite images and the relationship established by Schroeder et al. (2012)
between measured salinity and CDOM was used to estimate sea surface salinity in the flood plumes. All of the
regional pesticide maps were imported in ArcGIS for post-processing. Missing information (related to
atmospheric perturbations, cloud cover or reefs that were masked out) was interpolated in ArcGIS. Pesticide
levels were classified into different level of risk and the areas of reef and seagrass meadows at risk for each NRM
region were quantified.
Two different but complimentary methods were used to determine the risk posed by mixtures of PSII herbicides.
These were the Toxic Equivalence Quotient (TEQ) method (eg. Kennedy et al. 2012a; Smith et al. 2012) and the
multiple substances potentially affected fraction (ms-PAF) method (Traas et al, 2002). Importantly both methods
use the concentration addition model to determine the toxicity of mixtures of PSII herbicides. The maps shown
in this assessment are from the TEQ method.
Assessment classes:
The key PSII herbicides of concern (diuron, hexazinone, atrazine, tebuthiuron, ametryn and simazine) were
normalised to an herbicide-equivalent concentration which is based on the relative toxicity of diuron; the risk
Page 42
posed by PSII herbicides collectively could then be examined using the concentration addition model for joint
toxicity (see Kennedy et al. 2012a). The relative toxicities (EC50s and EC25s) of marine organisms including coral
species (Seriatopora hystrix and Acropora formosa), diatoms (Phaeodactylum tricornutum) and green algae
(Chlorella vulgaris) (Jones and Kerswell, 2003; Bengtson Nash et al. 2005; Muller et al. 2008) to each PSII
herbicide compared to diuron was determined and then averaged to produce the relative toxicity factors (RTFs)
(Kennedy et al. 2012a). The TEQ method was applied to the measured EC50s and EC25s of PSII herbicides that
inhibit the effective quantum yield (YII) in plants. Inhibition in YII by PSII herbicides is proportional to inhibition
of photosynthesis and growth in tropical microalgae (Magnusson et al. 2008) as well as reduced energy
acquisition by the host coral from its photosynthetic symbionts (Cantin et al. 2009).
Based on the toxicity of diuron calculated in several studies on coral and seagrass species we devised a set of
threshold values that were considered to match the following risk classifications:
Very High: >10 µg/L causes reduced growth and mortality in seagrass (Gao et al. 2011) and loss of
symbionts (bleaching) in corals (Jones et al. 2003; Negri et al. 2005). The effect on health and survival of
foundation species of the GBR can be catastrophic.
High: 2.3 – 10 µg/L Photosynthesis is reduced by between 50% and 90% in corals (Jones and Kerswell,
2003; Negri et al. 2011); seagrass (Chesworth et al. 2004; Gao et al. 2011; Flores et al. in review) and
microalgae (Magnusson et al. 2008, 2010). A 50% reduction of growth and biomass of tropical
microalgae was also reported in this concentration range (Magnusson et al. 2008). The community
structure of tropical microalgae is significantly affected and this causes significant changes in the
tolerance of microbial communities to herbicides (Magnusson et al. 2012). The effect on primary
production is major.
Medium: 0.5-2.3 µg/L Photosynthesis is reduced by between 10% and 50% in corals (Negri et al. 2011);
seagrass (Haynes et al. 2000; Chesworth et al. 2004; Gao et al. 2011; Flores et al. in review) and
microalgae (Magnusson et al. 2008, 2010). The community structure of tropical microalgae can be
affected by concentrations of diuron as low as 1.6 µg/L (Magnusson et al. 2012). The effect on primary
production is moderate.
Low: 0.1-0.5 µg/L Photosynthesis is reduced by up to 10% in corals (Negri et al. 2011); seagrass (Haynes
et al. 2000; Chesworth et al. 2004; Gao et al. 2011; Flores et al. in review) and microalgae (Magnusson et
al. 2008, 2010). The effect on primary production is minor.
Very Low: 0.025-0.1 µg/L No observed effect on photosynthesis in corals (Negri et al. 2011); seagrass
(Haynes et al. 2000; Flores et al. in review) and microalgae (Magnusson et al. 2008, 2010).
No Risk: < 0.025 µg/L
The highest risk classification determined for any point in a flood plume from a catchment was adopted as the
risk posed by that catchment.
Further explanations of the methods are provided in Chapter 4 of the Supporting Studies (Lewis et al. 2013a).
e) Recognising and assessing uncertainties in the selection of variables
For all variables, any relative differences in uncertainty and hence our confidence in the data can only be
assessed highly subjectively. If such qualitative assessments of uncertainty in our methodologies and data were
undertaken, uncertainty would be assessed as varying as much within as among NRM regions. As we compare
Page 43
results for NRM regions in the final combined relative risk assessment, the various methodologies used to
generate the data are considered to have roughly the same uncertainty and with the limited time and resources,
no specific estimates were considered.
Further discussion of the uncertainties and limitations of the assessment are presented in Part B of this report,
Section 4.
2.3.2 Estimating habitat area
The habitats considered in the assessment were for coral reefs and seagrass meadows, based on the best
available information. The area of GBR lagoon waters in each NRM region was also included in the assessments
as it contains other important habitats and biological populations such as phytoplankton, fish and benthic
organisms. The area estimates in each Region are based on the GBRMPA Spatial Data Centre’s coral reefs spatial
data file (December 2012) and seagrass spatial data files supplied by TropWATER James Cook University. The
seagrass habitat map used is comprised of a composite of the survey data up to 2010 (observed habitat) and a
statistical model of seagrass present in GBRWHA waters >15 metres depth. In this model spatial distribution is a
statistically modeled probability of seagrass presence (using generalised additive models with binomial error and
smoothed terms in relative distance across and along the GBR), based on ground truthed points (Coles et al.
2009). Locations with seagrass habitat probability >0.5 were included in the assessment.
2.3.3 Assessment Method - Part 1: Differential risk between pollutants on GBR ecosystems
The relative risk of different pollutants between regions was estimated by calculating the area of coral reefs,
seagrass meadows and GBR lagoon waters associated with an NRM region within each of the assessment classes
for each variable. As the first step, the assessment classes for each variable were allocated score in MCAS-S
between 0 (lowest severity) and 1 (highest severity) at the 1 km2 pixel scale. These are shown in Table 2.3 below.
Pixels in the highest assessment class all received the maximum value of 1. For example, for the TSS threshold of
2 mg/L the scores for the frequency of exceedance classes would be Very Low (<1% exceedance) = 0; Low (1-
10% exceedance) = 0.25; Medium (10-20% exceedance) = 0.5; High (20-50% exceedance) = 0.75; and Very High
(50-100% exceedance) = 1.0. The areas of coral reefs, seagrass meadows and GBR lagoon waters were then
reported for each assessment class in each region in ArcGIS.
To compare results between regions, only the areas affected in the highest assessment class for each variable
was considered as these were determined through expert opinion to be the most ecologically relevant in
determining risk, except for PSII herbicides where the highest and second highest assessment classes were
included in recognition of the toxicity of both of these classes. The output is a map and a table showing the area
(km2) of coral reef, seagrass and GBR lagoon waters within each assessment class for all variables in all NRM
regions. The assessment classes used in this part of the analysis are highlighted in grey in Table 2.3. For example,
the TSS threshold of 2 mg/L used in the assessment is the highest frequency of exceedance class of 50 to 100%
(Very High).
The results were then anchored for comparison; the maximum area is set as an anchor point and given a value
of 100, and all other area calculations are then expressed as a proportion of the maximum (values between 0
and 100). For example, for the TSS threshold of 2 mg/L the area of coral reef within the highest assessment class
(Very High – 50-100% exceedance of the threshold) is 9 km2 in the Burdekin, 6 km2 in the Fitzroy, and all other
regions have less than 2 km2 of coral reef in the area of the Very High assessment class. In this case the
maximum area of 9 km2 is in the Burdekin, so the Burdekin is allocated a value of 100. The other areas are then
reported as a proportion of the maximum value of 9 km2; the Fitzroy is therefore 61% and all other regions are
Page 44
less than 15%. It can then be concluded that the area of coral reefs in the Fitzroy region within the Very High
assessment class for TSS threshold of 2 mg/L is 61% of that in the Burdekin region.
Table 2.3. Summary of the classes used to assess the differential risk between pollutants on GBR ecosystems (Part 1), and
the weightings given to each assessment class for the MCAS-S combined assessment (Part 2). The cells shaded in grey
show the classes included in assessing the relative risk between variables (Part 1). The overall weighting for the water
quality variables used in the combined assessment are also shown. The variables are described in Table 2.1.
Variables
Overall
weighting
Assessment Class
Very Low
1
Low
2
Medium
3
High
4
Very High
5
TSS threshold exceedance 2mg/L
Frequency of exceedance (%)
<1
1-10
10-20
20-50
50-100
MCAS-S1 score
1/7
0
0.25
0.5
0.75
1.0
TSS threshold exceedance 7 mg/L
(5NTU)
Frequency of exceedance (%)
0
<1
1-10
10-20
20-50
50-100
MCAS-S score
1/7
0
0
0.33
0.66
1.0
TSS Plume Loading
(mean 2007-2011)
Category 1
Category 2
Category 3
MCAS-S score
1/7
0.33 2
0.66
1.0 3
Chl threshold exceedance
(0.45µg/L)
Frequency of exceedance (%)
<1
1-10
10-20
20-50
50-100
MCAS-S score
1/7
0
0.25
0.5
0.75
1.0
DIN Plume Loading
(mean 2007-2011)
Category 1
Category 2
Category 3
MCAS-S score
1/7
0.33 2
0.66
1.0 3
COTS Initiation Zone
Outside
Zone
Within Zone
MCAS-S score
1/7
0
1.0
PSII Herbicide modelled
concentration
(2009-2011) (µg/L)
0.025-0.1
0.1-0.5
0.5-2.3
2.3-10
>10
MCAS-S score
1/7
0.25
0.5
0.75
1.0
No
occurrence
1 MCAS-S: Multi-Criteria Analysis Shell for Spatial Decision Support developed by ABARES, refer to
http://www.daff.gov.au/abares/data/mcass; 2 This class covers Very Low and Low; 3 This class covers High and Very High.
2.3.4 Assessment Method - Part 2: Combined risk of degraded water quality to GBR ecosystems
To consider the combined risk of the selected water quality variables to marine ecosystems, the spatial layers
for the individual variables described above were combined and the scores allocated to each assessment class
for each variable were summed at the 1 km2 pixel scale (see Figure 2.7). Anchoring and normalising the data to
the uni-directional scale ensures that the ranges in data values (very different among variables) are standard
and can be summed to produce a single score for each pixel. Pixels with no data were not included in the final
averaging. It is recognised that this approach to standardising the assessment classes has limitations when
considering the equivalency of the assessment classes for each variable.
Page 45
Figure 2.7. Example of the results in one pixel from a Composite in MCAS-S. The result for the pixel from each layer is
summed to give a combined score. These scores are then classified into five assessment classes (Very Low to Very High).
In this example the combined score gives the pixel a score within the High assessment class in terms of relative risk of
degraded water quality.
The classifications, scores and overall weightings for this assessment were based on expert opinion and are
shown in Table 2.3. Ideally the classes for each variable would be scaled so that they are equivalent in terms of
potential ecological impacts to provide comparable weightings between variables. However, it is recognised that
this may not be the case for all variables given the inconsistencies in the temporal and spatial characteristics of
the datasets. As temporal and spatial resolution of data increases and the knowledge of the impacts of
sediments, nutrients and PSII herbicides on GBR ecosystems is advanced, this capability can be improved in
future assessments.
The data layers were then combined using the Composite tool in MCAS-S which essentially sums the results of
the scores for each 1 km2 pixel in each spatial layer to create a combined spatial layer. Figure 2.8 illustrates the
use of the MCAS-S tool. The result of the summed pixels can then be normalised in MCAS-S to generate a score
between 0 and 1 for each pixel (see example in Figure 2.9). The results for each pixel were further classified into
five even break classes ranging from Very Low to Very High to provide a classification of relative risk of degraded
water quality. A more detailed classification of 10 classes was also tested but the spread of the data resulted in
Page 46
limited benefit from using more than five classes. An example of the process applied in MCAS-S is shown in
Figure 2.7.
The area of coral reefs, seagrass meadows and GBR lagoon waters within each of the five assessment classes
was calculated in ArcGIS and tabulated for comparison between regions. A Marine Risk Index was defined by
summing the areas of coral reefs, seagrass and GBR lagoon waters in the Very High and High assessment classes
of the combined layer, and anchoring those results to the maximum area among regions. This enabled an
assessment of the relative differences between regions in terms of combined water quality risk for coral reefs,
seagrass and GBR lagoon waters. The Very High and High assessment classes were determined (by expert
opinion) to be the most ecologically relevant for the assessment. It is recognised that these classes are relative
and that the areas of coral reefs and seagrass meadows in the lower assessment classes may also be important
depending on temporal and spatial variability of the exposure to exceedance in the water quality variables. A
more detailed assessment of these patterns in the lower assessment classes was outside of the scope of this
project but should be considered in future work, particularly given the potential influence of chronic exposure to
pollutants, or the effects of periodic exposure to high concentrations of pollutants.
The final output is a Coral Reef Marine Risk Index and a Seagrass Marine Risk Index.
A number of options for combining the variables were considered which essentially weighted each variable
differently, described in Appendix 2. All of these were performed using the Composite tool in MCAS-S described
above. The selected option involved combining all variables individually with an equal weighting (1/7) and
summed. The maximum score for each pixel is therefore 7 (Figure 2.8), and this value is then normalised
between 0 and 1 (Figure 2.9). This approach essentially weights the variables differently because there are 3
sediment related variables, 3 nutrient related variables and 1 pesticide related variable which were considered
to be appropriate by the assessment team given current evidence of the relative importance of nutrients and
sediments compared to pesticides in the GBR. An example of this combination in MCAS-S is shown in Figure 2.7.
Page 47
Figure 2.8. Example of the results in one pixel (near the arrow)
from a Composite in MCAS-S where all individual variables are
combined equally (formula shown in the panel to the left). The
score for each variable is shown in the Viewer: the upper box
shows the combined score for each variable where variables
with no result or zero are allocated a value of -9999 (and not
counted), the lower box identifies the input scores for each
variable in the selected pixel (in this example the COTS
Initiation Zone did not receive a score as the pixel is outside the
Zone). The scores can then be normalised as shown in Figure
2.9.
Figure 2.9. Example of the results in one pixel (near the arrow)
from a normalised Composite in MCAS-S where all individual
variables are combined equally (shown in the panel to the left).
The classes and distribution of pixels within these classes for
the combined layer is shown at the bottom of the left panel.
This combined class for each variable is shown in the Viewer:
the upper box shows the combined score for the selected pixel,
the lower box identifies the scores for each input variable in the
selected pixel.
Page 48
2.3.5 Assessment Method - Part 3: Relative risk of degraded water quality to GBR ecosystems
To inform management priorities that aim to address the risks identified in Part 1 and 2, it is necessary to
understand the influence of river discharge in each of the regions, as these discharges carry a majority of the
pollutants into the GBR lagoon.
To relate the results of the Marine Risk Index (described in Section 2.3.4) to land based influences,
anthropogenic end-of-catchment loads were expressed as the proportion of total GBR load for each Region to
generate a Loads Index for each region. This recognises that while the total GBR load is important in influencing
the marine water quality conditions, it is only the anthropogenic proportion that can be factored into
management. The regional proportional contributions were then anchored (to normalise to a standard scale)
and averaged to generate a Loads Index for TSS, DIN and PSII herbicides for each NRM region. This assumes that
the relative importance of each load is equal which may not be the case, although there is currently insufficient
knowledge to weight the importance of the three pollutants relative to each other.
End-of-catchment anthropogenic loads were obtained from the results of the Source Catchments model
framework which have been produced as part of the Paddock to Reef Program (Waters et al. in press). First, the
Source Catchments modelling framework was used as a synthesis tool that incorporates new information on
paddock modelling of TSS, speciated N and P, and PSII herbicides, plus spatially and temporally remote sensed
inputs. This resulted in a consistent set of end of catchment pollutant loads for each of the 35 GBR catchments.
Anthropogenic load is calculated as the difference between the long term average annual load and the
estimated pre-European annual loads. A fixed climate period was used (1986 to 2009) for all model runs to
normalise for climate variability and provide a consistent representation of pre-development and anthropogenic
generated catchment loads. This therefore represents an ‘average’ year rather than the extremes such as those
recorded in the period 2008 to the current wet season in 2013. In addition, functionality from the previous
iteration catchment modelling, SedNet/ANNEX (for example see Cogle et al. 2006), was incorporated into Source
Catchments to represent hillslope, gully and streambank erosion and floodplain deposition processes.
It is recognised that assessment of the input of pesticides from each region can be expressed in a number of
ways, and while loads allow comparison between regions, it is the toxicity and therefore concentration that is
most relevant to the receiving environment. However, pesticide concentration data is currently limited across
the GBR. Therefore, in the final conclusions relating to pesticide risk in this assessment, additional evidence is
drawn from a combination of load and concentration data from specific locations, assessed in Lewis et al.
(2013a).
An index specific to the potential influence of river DIN loads on the initiation of COTS primary outbreaks in the
GBR was added (see Furnas et al. 2013a). DIN runoff is considered to be an important factor as approximately
40% of the loss of coral cover in the GBR since 1987 has been attributed to COTS predation (De’ath et al. 2012).
A COTS outbreak initiation zone has been defined between Lizard Island (14.5°S) and Cairns (17°S). On total
volumetric basis, most (86%) of the estimated freshwater input (direct and indirect) to the Zone comes from
Wet Tropics rivers, with the remaining 14% from the Burdekin River (Furnas et al. 2013a). These estimates were
used to create a COTS Influence Index.
To provide an overall relative ecological risk ranking between the NRM regions, the Marine Risk Indexes for coral
reefs and seagrass meadows were summed with the Loads Index, and for coral reefs only, the COTS Influence
Index, to generate a Coral Reef Relative Risk Index and a Seagrass Relative Risk Index. These final indexes for
coral reefs and seagrass were then summed and normalised to give an overall assessment of the relative risk of
degraded water quality to coral reefs and seagrass meadows to generate a Relative Risk Index for each region.
Page 49
Once the relative risks and pollutants are known, the load information for each catchment allows us to track
back to management priorities. A detailed assessment of the load contributions at a catchment scale was
outside of the scope of this assessment, however, the SCS chapters on sources of pollutants (Kroon et al. 2013)
and management practice effectiveness (Thorburn et al. 2013) provide a solid foundation for this analysis to be
progressed.
2.4 Results
2.4.1 Habitat and regional NRM areas
The distribution of coral reefs and seagrass used in the risk assessment are shown in Figure 2.10, and the inset
table shows the area of coral reef, seagrass and area of GBR lagoon waters in each marine NRM region. The total
area of the Great Barrier Reef Marine Park in this spatial dataset is 345,804 km2; this represents a difference of
well less than half of 1% the published area assessment from the GBRMPA of 345,400km2. The very slight
difference between the area estimates used here from ArcGIS and the published area assessments are a result
of slightly different estimates of the areas around coastlines and the Marine Park boundary. We consider the
difference here to be negligible.
The total area of coral reef in the GBR is estimated around 24,000 km2. The Regions with the largest areas of
coral reef are Cape York (10,354 km2), Fitzroy (4,855 km2), and Mackay Whitsunday (3,213 km2). Approximately
35,000km2 of seagrass has been mapped in the coastal waters around Queensland and Torres Strait since the
mid 1980s. Surveys and statistical modelling of seagrass in offshore waters deeper than 15m shows 37,454
square kilometres of the sea floor within the Great Barrier Reef World Heritage Area and Torres Strait has some
seagrass present making Queensland’s seagrass resources globally significant.
From the mapping data used in this assessment, the Cape York marine NRM region also has the highest area of
seagrass with 11,378km2. The regions with the second and third-highest seagrass area are the Burnett Mary
(6,330 km2) and the Burdekin (6,083 km2), respectively. The area of seagrass in the Mackay Whitsunday region is
relatively low compared to other Regions with only approximately 430 km2. Deepwater seagrasses are sparse in
the Mackay Whitsunday region, particularly south of Mackay, where tidal velocities are high and no major
deepwater seagrass meadows exist (Coles et al. 2009). High current stress, low Secchi readings and coarse
mobile sediments generally make this an unsuitable habitat for seagrass growth.
It is important to note that the habitats of the Burnett Mary region are under estimated in this assessment, as
the GBR Marine Park and World Heritage Area boundary does not include all of the habitat areas that would be
affected by the catchments of the Burnett Mary region. In particular, there is a large area of seagrass to the
south of the boundary in Hervey Bay which is known to provide important habitat, and foraging grounds, for
species that also inhabit the GBR Marine Park. The total area estimate for seagrass in the Burnett Mary region is
around 8,000km2 (McKenzie et al. 2010b). These limitations are discussed further in Part B Section 4.
2.4.2 Part 1: Relative importance of different pollutants to GBR ecosystems
The following section presents the results of the individual variables considered in this assessment (refer to
Section 2.3.3 for a description of the methods). This part of the risk assessment identifies the areas where each
water quality variable is considered to pose the greatest relative risk to coral reefs and seagrass between the
NRM regions. The output can be used to guide priorities for management of individual pollutants between NRM
regions.
Page 50
Figure 2.10. Locations of coral reefs and seagrass meadows used for the risk assessment. Coral reef outlines used are per
the GBRMPA Spatial Data Centre official reefs spatial data layer 2013. Seagrass areas are observed (composite of
surveyed data as at June 2010) and modelled deepwater seagrass habitat after Coles et al. (2009). Inset table shows the
area of coral reef, seagrass, those habitats combined and region for each marine NRM region.
Page 51
The area of coral reefs, seagrass and GBR lagoon waters is shown for all assessment classes for all variables in
this section. Area calculations are rounded to the nearest whole km2 for ease of reporting but does include
summed portions of some 1 km2 pixels. For each variable there is a table of results, and a map shows the areas
within each assessment class for each variable. Within the map, a graph compares the areas of coral reef and
seagrass affected by the highest class and a comparison of these areas with the total (whole of GBR) areas of
habitat affected among the NRM regions is presented as a pie chart.
Note that in this section the regions are simply referred to as their name, for example ‘the Burdekin’ rather than
‘the Burdekin region’ to improve readability.
a) Sediments
Total suspended solids threshold exceedance, Threshold a – 2 mg/L
As shown in Table 2.1, five assessment classes were used for TSS 2 mg/L based on the frequency of exceedance
of this concentration (in days) in the period 2002 to 2012, expressed as a percentage of the total number of valid
daily observations ranging from Very Low to Very High. However, as the Very Low class receives a score of 0, it is
not reported here. The results of the assessment are shown in Table 2.4 and Figure 2.11. The areas within the
Very High class are constrained to the coast and concentrated in the Burdekin and Fitzroy. These inshore areas
are locations with some of the highest use and visitation rates; this is a result common to all individual variables
and is reviewed in the discussion.
Key findings:
The area of coral reef in the Very High class of TSS exceedance at 2 mg/L is greatest in the Burdekin (9
km2), and second-greatest in the Fitzroy (6 km2) (see inset in Figure 2.11). The greatest area within the
High class is in the Fitzroy (103 km2) and Mackay Whitsunday is second (90 km2). The proportion of coral
reefs in each region in the Very High class is less than 1%. More than 95% of the coral reefs within in
each region are within the lowest classes (Very Low and Low) for exceedance of TSS 2 mg/L.
The area of seagrass within in the Very High class of TSS exceedance at 2 mg/L is also greatest in the
Burdekin (209 km2), which is ~10 times greater than the second-greatest area in the Wet Tropics (22
km2) (see inset Figure 2.11). In the High class Cape York has the greatest area of seagrass (747 km2), with
Fitzroy second (300 km2). The proportion of seagrass meadows in each region in the Very High class is
less than 5%. More than 59% of the seagrass meadows within in each region are within the combined
Very Low and Low class for exceedance of TSS 2 mg/L and a large proportion of these are deepwater
meadows (>15m). The proportion of seagrass within all other categories is similar across regions (all less
than 10%) with the notable exception of Mackay Whitsunday which has 18% and 16% within the
Medium and High classes respectively.
The pie chart insets in Figure 2.11 show the area of coral reef and seagrass within each region as a
percentage of total area (GBR-wide) of coral reef and seagrass within the Very High class of TSS
exceedance at 2 mg/L. The area of coral reef in the Very High class in Burdekin represents 52% of the
total GBR coral reef area within the Very High class, and the area of seagrass in the Very High class in the
Burdekin represents 78% of the total GBR seagrass area within the Very High class. The lowest areas of
coral reefs and seagrass in both the Very High and High classes are in the Burnett Mary (<5 km2 for coral
and 55km2 for seagrass).
The area of GBR lagoon waters in the Very High class of TSS exceedance at 2 mg/L is greatest in the
Burdekin (932 km2), and second in the Fitzroy (502 km2). The greatest area in the High class is in the
Page 52
Fitzroy (3,390 km2) with Mackay Whitsunday second (2,122 km2). The proportion of GBR lagoon waters
in each region in the Very High class is less than 5%.
When summed, the area of coral reef within the High and Very High classes (frequency of exceedance
20-50% and 50-100%) of TSS exceedance at 2 mg/L represents 0.8% (Cape York) to 2.8% (Mackay
Whitsunday) of the total coral reef area in the regions; seagrass ranges from 0.9% (Burnett-Mary) to
17% (Mackay Whitsunday) of the seagrass area in the regions, and GBR lagoon waters from 1% (Burnett-
Mary) to 5% (Burdekin) of the total region area. The lowest total areas in the Very High and High classes
are in Burnett Mary, however the assessment area is bound to the GBRWHA and therefore does not
incorporate coral reefs that have not been fully mapped within the region, or the large seagrass
meadows in Hervey Bay the large seagrass meadows in Hervey Bay.
Table 2.4. Area of coral reefs, seagrass meadows and GBR lagoon waters within the Low to Very High assessment classes
for TSS 2 mg/L and the percent of the NRM region that the area represents. Results for the assessment are based on
frequency of exceedance of TSS 2 mg/L using daily remote sensing data 2002-2012 (see methods in Section 2.3.1a(i)).
TSS 2 mg/L
Low
Medium
High
Very High
Frequency of
exceedance class
1-10%
10-20%
20-50%
50-100%
NRM Regions
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Coral Reefs
Cape York
10,163
98
71
1
84
1
0
0
Wet Tropics
2,355
97
17
1
46
2
0
0
Burdekin
2,930
99
2
<1
9
<1
9
<1
Mackay Whitsunday
2,886
90
191
6
90
3
1
<1
Fitzroy
4,559
94
173
4
103
2
6
<1
Burnett-Mary
277
98
1
<1
3
1
2
1
Seagrass
Cape York
8,949
79
463
4
747
7
0
0
Wet Tropics
4,450
91
57
1
117
2
22
<1
Burdekin
4,289
71
54
1
165
3
209
3
Mackay Whitsunday
255
59
79
18
70
16
4
1
Fitzroy
5,042
87
159
3
300
5
11
<1
Burnett-Mary
5,926
94
24
<1
35
1
20
<1
GBR lagoon waters
Cape York
92,137
96
2,016
2
1,809
2
0
0
Wet Tropics
28,647
91
1,238
4
1,446
5
147
<1
Burdekin
43,582
93
734
2
1,497
3
932
2
Mackay Whitsunday
43,948
91
2,274
5
2,122
4
131
<1
Fitzroy
79,095
92
2,440
3
3,390
4
502
1
Burnett-Mary
36,514
98
412
1
233
1
145
<1
The results presented here and used in the risk assessment have been compared with spatial patterns in secchi
depth shown in De’ath and Fabricius (2008). The visual comparison of our outputs for TSS exceedance at 2 mg/L
and those from De’ath and Fabricius (2008) are shown in Appendix 1 along with a similar comparison for Chl
0.45 µg/L. The spatial patterns are very similar suggesting the remotely sensed data used here are at least
comparable to the long term patterns measured in situ.
Page 53
Figure 2.11. Results for the assessment of frequency of exceedance of TSS 2 mg/L using daily remote sensing data 2002-
2012. Results for the assessment are based on frequency of exceedance of TSS 2 mg/L (see methods in Section 2.3.1a(i)).
Inset bar chart compares coral reef and seagrass area within the Very High class (50-100% exceedance); pie charts show
the area of coral reef and seagrass within each region as a percentage of total area (GBR-wide) of coral reef and seagrass
within the Very High class.
Page 54
Total suspended solids threshold exceedance, Threshold b - 7 mg/L (turbidity 5NTU)
As shown in Table 2.1, five assessment classes were used for TSS 7 mg/L (5NTU) based on the frequency of
exceedance of this concentration (in days) in the period 2002 to 2012, expressed as a percentage of the total
number of valid daily observations ranging from Very Low to Very High, but only the results of the Medium to
Very High classes are presented as the most relevant here. Note that the assessment classes are different from
those for TSS 2 mg/L to reflect the greater severity of the higher concentration; however, there were no pixels
where the frequency of exceedance was greater than 50%. The results of the assessment are shown in Table 2.5
and Figure 2.12. The area exposed in the Very High class is constrained to the coast (see map forming Figure
2.12) and concentrated in the Fitzroy region.
Table 2.5. Area of coral reefs, seagrass meadows and GBR lagoon waters within the Medium to Very High assessment
classes for TSS 7 mg/L (5 NTU) and the percent of the NRM region that the area represents. Results for the assessment
are based on frequency of exceedance of TSS 7 mg/L (5 NTU) using daily remote sensing data 2002-2012 (see methods in
Section 2.3.1a(ii)).
TSS 7 mg/L
Medium
High
Very High
Frequency of
exceedance class
1-10%
10-20%
20-100%
NRM Regions
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Coral Reefs
Cape York
1,280
12
3
<1
0
0
Wet Tropics
423
17
0
0
0
0
Burdekin
105
4
<1
<1
<1
<1
Mackay Whitsunday
747
23
5
<1
0
0
Fitzroy
775
16
25
1
11
<1
Burnett Mary
58
20
0
0
0
0
Seagrass
Cape York
2,437
21
30
<1
0
0
Wet Tropics
624
13
50
1
9
<1
Burdekin
600
10
96
2
35
1
Mackay Whitsunday
219
51
16
4
0
0
Fitzroy
484
8
29
<1
9
<1
Burnett Mary
393
6
6
<1
0
0
GBR lagoon waters
Cape York
9,661
10
80
<1
0
0
Wet Tropics
5,063
16
204
1
30
<1
Burdekin
7,381
16
451
1
170
<1
Mackay Whitsunday
6,026
12
185
<1
17
<1
Fitzroy
9,329
11
1,094
1
504
1
Burnett Mary
2,062
6
26
<1
0
0
Key findings:
The area of coral reef in the Very High class of TSS exceedance at 7 mg/L is greatest in the Fitzroy (11
km2); all other regions have less than 1 km2 or zero (see inset in Figure 2.8). The greatest area within the
High class is in the Fitzroy (25 km2) and Mackay Whitsunday is second (5 km2). The proportion of coral
Page 55
reefs in each region in the High and Very High class is less than 1%. More than 80% of the coral reefs
within each region are within the lowest classes where the frequency of exceedance of TSS 7 mg/L was
<1% (not shown here). The lowest areas of coral reefs in all assessment classes are in Burnett Mary.
The area of seagrass within in the Very High class of TSS exceedance at 7 mg/L is also greatest in the
Burdekin (35 km2), with 9 km2 within in both the Wet Tropics and Fitzroy (see inset Figure 2.12). In the
High class Burdekin has the greatest area of seagrass (96 km2), with Wet Tropics second (50 km2). The
proportion of seagrass meadows in each region in the High class is less than 5%, and less than 1% in the
Very High class. More than 80% of the seagrass meadows within all regions except Mackay Whitsunday
are within the lowest classes where the frequency of exceedance of TSS 7 mg/L was <1% (not shown
here). In Mackay Whitsunday 55% of the seagrass is within the Medium to Very High classes.
The area of GBR lagoon waters in the Very High class of TSS exceedance at 7 mg/L is greatest in the
Fitzroy (504 km2), and second in the Burdekin (170 km2). The greatest area in the High class is in the
Fitzroy (1,094 km2) with Mackay Burdekin second (451 km2). The proportion of GBR lagoon waters in
each region in the Very High class is less than 1%. More than 80% of the GBR lagoon waters within in
each region are within the lowest classes where the frequency of exceedance of TSS 7 mg/L was <1%
(not shown here).
The pie chart insets in Figure 2.12 show the area of coral reef and seagrass within each region as a
percentage of total area (GBR-wide) of coral reef and seagrass within the Very High class of TSS
exceedance at 7 mg/. The area of coral reef in the Very High class in Fitzroy represents 97% of the total
GBR coral reef area within the Very High class, and the area of seagrass in the Very High class in the
Burdekin represents 65% of the total GBR seagrass area within the Very High class.
When summed the area of coral reef within the High and Very High of TSS exceedance at 7 mg/L
represents 0% (Wet Tropics) to 0.7% (Fitzroy) of the total coral reef area in the regions; seagrass ranges
from 0.1% (Burnett Mary) to 3.7% (Mackay Whitsunday) of the total seagrass area in the regions, and
GBR lagoon waters from 0.1% (Burnett Mary) to 1.9% (Fitzroy) of the total region area. The total areas in
the High and Very High classes are lowest in Burnett Mary; however the assessment area is bound to the
GBRWHA and therefore does not incorporate coral reefs that have not been fully mapped within the
region, or the large seagrass meadows in Hervey Bay.
For both concentrations of TSS exceedance, there is a distinct area of exceedance of the Medium class in the
coastal areas in the Cape York region north of Cooktown and Princess Charlotte Bay. Further validation of this
result is required, however recent studies by Brooks et al. (2013) indicate that suspended sediment loads from
the Normanby River are likely influence this area. Similar patterns exist in the coastal areas around Shoalwater
Bay in the northern part of the Fitzroy region which are also known to be naturally turbid and uncertainties in
the remote sensing results in these areas have not been resolved. While these uncertainties may be important
for regionally specific analyses, they are not considered to be significant enough to influence the overall
conclusions of this assessment.
Page 56
Figure 2.12. Results for the assessment of frequency of exceedance of TSS 7 mg/L (5NTU) using daily remote sensing data
2002-2012. Results for the assessment are based on frequency of exceedance of TSS 7 mg/L (5NTU) (see methods in
Section 2.3.1a(i)). Inset bar chart compares coral reef and seagrass area within the Very High class (50-100% exceedance);
pie charts show the area of coral reef and seagrass within each region as a percentage of total area (GBR-wide) of coral
reef and seagrass within the Very High class.
Page 57
TSS plume loading (mean 2007-2011)
As shown in Table 2.1, three assessment classes were used for TSS plume loading based on plume frequency
information from remote sensing and scaled river load data. To factor in inter-annual variability, we used the
mean result of assessments completed annually from 2007 to 2011. The results of the assessment are shown in
Table 2.6 and Figure 2.13. Note that there are areas within the GBRWHA that are assessed as being outside of
the area of plume loading influence which are not reported in this assessment. The areas within the highest
assessment class (High) are inshore (Figure 2.13) and mostly concentrated in the Burdekin region.
Table 2.6. Area of coral reefs, seagrass and GBR lagoon waters within the assessment classes for TSS mean plume
loadings between 2007 and 2011, and the percent of the NRM region that the area represents. The assessment classes
are relative and derived from a combination of scaled river loads data and flood plume frequency analysis from remote
sensing data (see methods in Section 2.3.1b).
TSS plume loadings
Low
Medium
High
NRM Regions
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Coral Reefs
Cape York
8,903
86
0
0
0
0
Wet Tropics
2,267
93
23
1
2
<1
Burdekin
2,746
93
35
1
20
1
Mackay Whitsunday
401
12
229
7
1
0
Fitzroy
694
14
161
3
6
0
Burnett Mary
268
95
3
1
0
0
Seagrass
Cape York
11,433
100
0
0
0
0
Wet Tropics
4,535
93
223
5
48
1
Burdekin
5,113
84
354
6
610
10
Mackay Whitsunday
<1
<1
407
95
4
1
Fitzroy
5,244
91
490
8
30
1
Burnett Mary
5,750
91
341
5
0
0
GBR lagoon waters
Cape York
54,536
57
0
0
0
0
Wet Tropics
21,464
68
2,817
9
445
1
Burdekin
21,899
47
7,160
15
4,553
10
Mackay Whitsunday
13,004
27
12,331
25
136
0
Fitzroy
26,488
31
13,506
16
763
1
Burnett Mary
10,172
27
1,802
5
0
0
Key findings:
The area of coral reef in the High class for TSS plume loading is greatest in the Burdekin (20 km2), and
second-greatest in the Fitzroy (6 km2) (see inset in Figure 2.13). The greatest area within the Medium
class is in Mackay Whitsunday (229 km2) and Fitzroy is second (161 km2). The proportion of coral reefs in
each region in the High class is less than 1%. More than 85% of the coral reef area in the Cape York, Wet
Tropics and Burdekin are within the Low class, and approximately 80% of the area of coral reefs in the
Mackay Whitsunday and Fitzroy are outside the mapped plume loading area. There is no occurrence of
coral reefs in the High class in Cape York and Burnett Mary, but there are limitations to the method
Page 58
applied to for Cape York described in Section 2.3.1c, and the area of coral reefs in the Burnett Mary is
not considered to be accurate.
The area of seagrass within the High class for TSS plume loading is also greatest in the Burdekin (610
km2), which is considerably larger than the second-greatest area in the Wet Tropics (48 km2) (see inset
Figure 2.13). In the Medium class Fitzroy has the greatest area of seagrass (490 km2) with Mackay
Whitsunday second (407 km2). The proportion of seagrass meadows in each region within the High class
is less than 1%, except for Burdekin which is 10%. Approximately 95% of the seagrass area in Mackay
Whitsunday region is in the Medium class, and all other regions are less than 10%, which is associated
with the distribution of seagrass in the region which is predominantly inshore (see 2.4.1 for further
explanation). For all other regions, more than 80% of the seagrass in the region is within the Low class.
There is no occurrence of seagrass meadows in the High class in Cape York and Burnett Mary, but there
are limitations to the method applied to for Cape York described in Section 2.3.1c, and large seagrass
meadows exist outside of the GBRWHA boundary and were not included in this assessment.
The pie chart insets in Figure 2.13 show the area of coral reef and seagrass within each region as a
percentage of total area (GBR-wide) of coral reef and seagrass within the High class for TSS plume. The
area of coral reef and seagrass in the Burdekin represents 68% and 88% of the total GBR coral reef area
and seagrass area respectively of habitats within the High class across the GBR.
The total extent of influence of all TSS plume loading classes varies between the regions but is greater
than 50% of the total region area in all cases except for Burnett Mary which is 32%. Large proportions
(>70%) of the Wet Tropics and Burdekin are within the total area of TSS exposure. The area of GBR
lagoon waters in the High class is greatest in the Burdekin (4,553 km2), and second in the Fitzroy (763
km2), associated with large river discharges in the assessment period (Figure 2.13). The greatest area in
the Medium class is in the Fitzroy (13,506 km2) with Mackay Whitsunday second (12,331 km2). The
proportion of GBR lagoon waters in each region in the High class is less than 1%, except for the Burdekin
which is 10%. In Cape York, 43% of the area is outside of the estimated plume loading area, although
there are obvious uncertainties in the method used for this Region which must be taken into account
(see Section 2.3.1c).
When summed the coral reef area in the Medium and High classes for TSS plume loading represents 0%
(Cape York) to 7.1% (Mackay Whitsunday) of the total coral reef area in the regions; seagrass ranges
from 0% (Cape York) to 95% (Mackay Whitsunday) of the total seagrass area in the regions, and GBR
lagoon waters from 0% (Cape York) to 26% (Mackay Whitsunday) of the total region area. The areas of
lowest exposure are within the Cape York and Burnett Mary regions, although these results cannot be
concluded with a great degree of certainty due to limited or no validation in those locations.
Page 59
Figure 2.13. Results for the assessment of TSS plume loading (mean of annual assessments 2007 to 2011). The
assessment classes are relative and derived from a combination of scaled river loads data and flood plume frequency
analysis from remote sensing data (see methods in Section 2.3.1c). Inset bar chart compares coral reef and seagrass area
within the High class; pie charts show the area of coral reef and seagrass within each region as a percentage of total area
(GBR-wide) of coral reef and seagrass within the High class.
Page 60
b) Nutrients
Chlorophyll threshold exceedance 0.45 µg/L
Chlorophyll (Chl) concentrations are relevant year round as an indication of nutrient enrichment in marine
waters. As shown in Table 2.1, five assessment classes were used for Chl 0.45 µg/L based on the frequency of
exceedance of this concentration (in days) in the period 2002 to 2012, expressed as a percentage of the total
number of valid daily observations ranging from Very Low to Very High. However, as the Very Low class receives
a score of 0, it is not reported here.The results of the assessment are shown in Table 2.7 and Figure 2.14. The
areas within the Very High class are constrained to the coast and concentrated in the Burdekin and Fitzroy, as
was the case for TSS 2 mg/L (Section 2.4.2a). These inshore areas are locations with some of the highest use and
visitation rates; this is a result common to all individual variables and is reviewed in the discussion.
Table 2.7. Area of coral reefs, seagrass meadows and GBR lagoon waters within the Low to Very High assessment classes
for Chl 0.45 µg/L and the percent of the NRM region that the area represents. Results for the assessment are based on
frequency of exceedance of Chl 0.45 µg/L using daily remote sensing data 2002-2012 (see methods in Section 2.3.1b).
Chl 0.45 µg/L
Low
Medium
High
Very High
Frequency of
exceedance class
1-10%
10-20%
20-50%
50-100%
NRM Regions
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Area
(km2)
% of
habitat in
region
Coral Reefs
Cape York
8,364
81
1,324
13
614
6
<1
<1
Wet Tropics
2,063
85
288
12
65
3
<1
<1
Burdekin
2,623
88
256
9
34
1
12
<1
Mackay Whitsunday
707
22
1,599
50
846
26
10
<1
Fitzroy
2,463
51
1,620
33
707
15
46
1
Burnett-Mary
216
76
54
19
8
3
4
1
Seagrass
Cape York
6,527
57
2,410
21
2,410
21
14
<1
Wet Tropics
4,055
83
554
11
204
4
45
1
Burdekin
5,435
89
79
1
231
4
315
5
Mackay Whitsunday
46
11
114
27
223
52
23
5
Fitzroy
5,042
87
291
5
230
4
196
3
Burnett-Mary
5,942
94
217
3
110
2
47
1
GBR lagoon waters
Cape York
47,149
49
7,632
8
6,225
6
24
<1
Wet Tropics
17,750
56
2,033
6
3,218
10
617
2
Burdekin
29,989
64
2,072
4
2,439
5
2,202
5
Mackay Whitsunday
26,845
55
10,641
22
7,431
15
632
1
Fitzroy
63,314
74
9,886
12
7,601
9
2,140
3
Burnett-Mary
34,563
93
1,013
3
829
2
419
1
Page 61
Key findings:
The area of coral reef in the Very High class of Chl exceedance of 0.45 µg/L is greatest in the Fitzroy (46
km2), and second-greatest in the Burdekin (12 km2) (see inset in Figure 2.14). The greatest area within
the High class is in Mackay Whitsunday (846 km2) and Fitzroy is second (707 km2). The proportion of
coral reefs in each region in the Very High class is less than 1%. More than 75% of the coral reefs in Cape
York, Wet Tropics, Burdekin and Burnett Mary are within the Low class, and 50% of the Fitzroy is within
the Low class of frequency of exceedance of 0.45 µg/L.
The area of seagrass within in the Very High class Chl exceedance of 0.45 µg/L is greatest in the Burdekin
(315 km2), and the second-greatest area is the Fitzroy (196 km2) (see inset Figure 2.10). In the High class
Cape York has the greatest area of seagrass (2,410 km2) with the Wet Tropics, Burdekin, Mackay
Whitsunday and Fitzroy all similar with 200-230 km2 of seagrass in the High class of Chl 0.45 µg/L
exceedance. The proportion of seagrass meadows in each region in the Very High class is less than 5%.
More than 80% of the seagrass meadows within the Wet Tropics, Burdekin, Fitzroy and Burnett Mary